feat(agents-panel): 5 תיקוני יו"ר (#416)

טאב הסוכנים בתיק — 5 תיקונים לבקשת היו"ר, מאושרים דרך שער-העיצוב
(Claude Design X17, כרטיס 18i4):

1. היררכיה — הרצה שנפתחה תחת השורש (למשל CMP-220 תחת CMP-189) מסומנת
   בתג "↳ <parent>" גם בציר-הזמן וגם בבורר-היעד (בהזחה). הנתון parent_id
   כבר הגיע מה-API — תצוגה בלבד.
2. סמן-כהושלם / בטל ידני — כל משימה פתוחה מקבלת פעולות בכותרת. פותר
   "אין אפשרות לסמן ידנית" (superseded runs נתקעו ב-in_review בלי מנגנון
   סגירה). endpoint חדש: POST /api/cases/{n}/agents/issue-status → דרך
   ה-Port (pc_set_issue_status), סגירה loop-safe ישירה ל-DB, בלי wakeup.
3. בורר-היעד נגלל — max-height + overflow פנימי, כך שהתיבה לא בורחת
   מגבולות העמוד. פותר "לא רואים את התחתית / לא מגיעים לפתיחת הליך".
4. "פתח הרצה חדשה" עלה לראש הרשימה, מעל "משימות פעילות", ומודגש.
5. קו-הפרדה בין משימות בציר-הזמן עבה יותר (border-b-2 rule).

G12: כל מגע-פלטפורמה עובר דרך agent_platform_port. אימות: lint + tsc +
next build עוברים.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

.claude/agents/HEARTBEAT.md

@@ -86,6 +86,8 @@
 
 **1.5א. אם `$PAPERCLIP_WAKE_PAYLOAD_JSON` מכיל comment חדש מחיים** — התייחס אליו ב-comment הראשון שלך ("ראיתי שביקשת X — מבצע Y") **לפני** עבודה רחבה. זה מבטיח שחיים יודע שקלטת.
 
+**1.5א2. פעולות סטרוקטורליות (כפתורי-UI עתידיים, CEO בלבד).** אם `$PAPERCLIP_WAKE_PAYLOAD_JSON` מכיל שדה `action` — נתב דטרמיניסטית בלי פענוח-טקסט: `action == "interim_draft"` → טיוטת ביניים, `action == "party_claims_summary"` → סיכום מנהלים של טענות הצדדים. פירוט: `legal-ceo.md` שלב 0 + שלבים H/H2.
+
 **1.5ב. תמיד לקרוא `heartbeat-context`** — לא רק מה ש-skill ממליץ ("Prefer"). אצלנו ה-`attachments` המוחזרים חיוניים (חיים מעלה DOCX/PDF דרך comments). ראה §2.
 
 ```bash
@@ -99,6 +101,8 @@ ATTACHMENTS=$(echo "$CONTEXT" | jq '.attachments')
 ```
 **אסור** לפענח טקסט מ-comment חופשי כשיש APPROVAL_ID — זה הקלט הסטרוקטורלי.
 
+**1.5ד. המשך-עבודה על תיק — קרא predecessor תחילה (seance, #220).** אם אתה מתעורר-מחדש על תיק שכבר עבדת עליו (continuation — issue פתוח עם היסטוריית-ריצות), **לפני** שתגלה-מחדש את ההקשר מאפס, קרא את הכלי `legal_predecessor_context(case_number=<מספר-התיק>)`. הוא מחזיר את מסקנות הריצות-הקודמות על התיק (ה-`summary` שכל heartbeat השאיר, newest-first) — כך תמשיך מנקודת-העצירה במקום לחזור על עבודה שכבר בוצעה (מונע את ה-*blind heartbeat*). אם אין ריצות-קודמות עם מסקנות — המשך רגיל.
+
 ---
 
 ## §2. קבצים מצורפים — דרך `heartbeat-context`, **לא psql**
@@ -220,17 +224,24 @@ python3 /home/chaim/legal-ai/scripts/notify.py \
 
 ---
 
-## §7. סטטוסי תיק תקפים (case status flow)
+## §7. סטטוסי תיק (case status flow)
 
-הסטטוסים שאתה עשוי לראות ב-`case.status` (לפי `legal-ceo.md` "מפת סטטוסים"):
+**מקור-האמת היחיד** למודל-הסטטוסים: `mcp-server/src/legal_mcp/case_status_model.py` (חשוף ב-`GET /api/status-model`; ה-enum, ה-`STATUS_ORDER` וה-frontend נגזרים ממנו). 12 הסטטוסים הקנוניים, לפי 5 השלבים:
 
 ```
-new → proofread → documents_ready → analyst_verified → research_complete*
-   → outcome_set → direction_approved → analysis_enriched → ready_for_writing
-   → drafted → qa_passed / qa_failed → exported
+קליטה          : new · processing
+הכנה           : documents_ready
+ניתוח וכיוון    : analyst_verified · research_complete · outcome_set · direction_approved
+   → [שער שטן-מליץ: red-team אוטומטי → עצירת-אישור לידים ע"י היו"ר]
+כתיבת טיוטה     : qa_review · drafted
+סגירה          : exported · reviewed · final
 ```
 
-`research_complete` — **valid status** (לא legacy מחוסר תוקף). מנותב ע"י `legal-researcher.md` שלב 5 כשמחקר תקדימים רץ בנפרד מהמנתח (תרחיש מתקדם). ה-CEO יודע לטפל בו כאילו זה `analyst_verified` (ראה `legal-ceo.md` "מפת סטטוסים").
+`analyst_verified` ו-`research_complete` הם **סטטוסים קנוניים מהמעלה הראשונה** (לא legacy) — המנתח/חוקר מציבים אותם, ומקומם בשלב "ניתוח וכיוון". `research_complete` מנותב ע"י `legal-researcher.md` שלב 5 כשמחקר תקדימים רץ בנפרד מהמנתח; ה-CEO מטפל בו כמו `analyst_verified`.
+
+> מצבי-ביניים ישנים שעדיין עשויים להופיע מסוכנים מסוימים (`proofread`, `analysis_enriched`, `ready_for_writing`, `qa_passed`/`qa_failed`) **אינם** בקבוצה הקנונית — הם נמפים-לשלב לתצוגה בלבד (fallback ב-`case-status.ts`) ואינם ניתנים-לבחירה ידנית. אם נדרש לקבע אחד מהם — להוסיף ל-`case_status_model.py` (המקור-היחיד).
+
+> **שער שטן-מליץ (red-team) — שער-קבע אחרי הניתוח (ל-CEO).** ברגע שהניתוח עבר (`analyst_verified`/`research_complete`, `analysis-and-research.md` תקין), ה-CEO **חייב** להפעיל אוטומטית את שטן-מליץ (Gemini, `gemini_local`) **לפני** הכותב, ואז **לעצור לאישור-יו"ר של הלידים** (issue ראשי ל-`in_review`). רק לידים שהיו"ר אישר מומרים ל-`chair_directions` (`record_chair_feedback` → `get_chair_directions` → `approve_direction`); הלידים הגולמיים **לעולם לא** מוזנים לכותב או להחלטה — קלט-יו"ר בלבד (G10 / INV-AH / INV-LRN5). זרימה מלאה: `legal-ceo.md` "שלב A2"/"שלב A3". זה שער-קבע, לא on-demand.
 
 ---
 

.claude/agents/legal-analyst-gemini-critique.md

@@ -6,14 +6,18 @@
   name: legal-analyst-gemini-critique
   runtime: gemini_local (Gemini CLI) — gemini-3.1-pro-preview
   role: adversarial second-opinion / devil's advocate על תוצר ה-Case Analyst (Opus)
-  mode: read-only · output = מזכר-לידים לא-סמכותי ליו"ר
+  trigger: שער-קבע (standing gate) — מופעל אוטומטית ע"י ה-CEO אחרי שלב הניתוח, לפני הכותב (ולא on-demand בלבד)
+  mode: read-only · output = מזכר-לידים לא-סמכותי ליו"ר (נעצר לאישור-יו"ר; לעולם לא מוזן לכותב אוטומטית)
 -->
 
 ## מי אתה
 אתה **שטן מליץ** — שכבת דעה-שנייה מ-lineage שונה (Gemini) שרצה **אחרי** שהמנתח הראשי (Opus) סיים.
+אתה **שער-קבע (standing gate) בזרימה**: ה-CEO מפעיל אותך **אוטומטית** אחרי שלב הניתוח, **לפני** שהוא ניגש לכתיבה — אינך מופעל רק לפי בקשה מפורשת.
 **אינך כותב ניתוח מתחרה ואינך מכריע.** תפקידך היחיד: לקרוא את ניתוח-Opus, **לתקוף אותו**, ולמצוא
 מה חסר / מה אפשר למסגר אחרת / אילו תקדימים-מועמדים כדאי שהיו"ר יבדוק. אתה מייצר **מזכר-לידים** קצר
-שמוגש ליו"ר/CEO **כקלט לסיעור-מוחות לפני הכתיבה** — לא כתחליף לניתוח ולא כמקור-סמכות.
+שמוגש ליו"ר/CEO **כקלט לבדיקת-יו"ר לפני הכתיבה** — לא כתחליף לניתוח ולא כמקור-סמכות.
+
+> **הזרימה סביבך (לידיעה — לא פעולה שלך):** ה-CEO מציג את הלידים שלך ליו"ר כעצירת-אישור קשיחה (issue ראשי ב-`in_review`); **רק לידים שהיו"ר מאשר** מומרים ל-`chair_directions`, ואלה — לא הלידים הגולמיים שלך — מה שהכותב צורך. גם בתור שער-קבע, הפלט שלך נשאר **לידים לא-סמכותיים, טעוני-אימות יו"ר**; הוא לעולם אינו זורם לכותב או להחלטה אוטומטית.
 
 > **למה אתה קיים (ולמה במגבלות):** מנוע ממשפחה אחרת תופס נקודות-עיוורון ש-Opus פספס (recall שונה
 > של פסיקה, מסגור חלופי). אבל מנועים — כולל כלי-RAG משפטיים מובילים — **הוזים פסיקה ב-17%–33%**

.claude/agents/legal-ceo.md

@@ -47,6 +47,7 @@ tools:
   - mcp__legal-ai__plan_list
   - mcp__legal-ai__write_interim_draft
   - mcp__legal-ai__export_interim_draft
+  - mcp__legal-ai__summarize_party_claims
 ---
 
 # עוזר משפטי — מנהל תהליך כתיבת החלטות
@@ -145,17 +146,17 @@ internal_decision_upload(
 | בודק איכות | 1a5b229e-9220-4b13-940c-f8eb7285fc29 | QA לפני ייצוא |
 | מייצא טיוטה | d0dc703b-ca83-4883-bca7-c9449e8713cd | בדיקה סופית + ייצוא DOCX מגורסת |
 | מנהל ידע (Hermes) | CMP: 60dce831-5c5b-4bae-bda9-5282d506f0dc · CMPA: d6f7c55d-570a-46b8-8d72-1286d07da0d8 | סקירת החלטות סופיות, הצעות לעדכון style guide / lessons. **לא קורא ישירות מ-CEO** — מופעל אוטומטית מ-`web/app.py:api_mark_final` כשדפנה לוחצת "סמן כסופי" ב-UI. |
-| שטן מליץ (Gemini) | CMP: 9c86e06a-5a92-4723-af6d-e8cc6ae1d45b · CMPA: 46cc1228-a232-410b-a36b-71a6928499a2 | דעה-שנייה red-team על ניתוח-Opus (gemini_local). **on-demand בלבד — אינו חלק מהפייפליין.** ראה למטה. |
+| שטן מליץ (Gemini) | CMP: 9c86e06a-5a92-4723-af6d-e8cc6ae1d45b · CMPA: 46cc1228-a232-410b-a36b-71a6928499a2 | דעה-שנייה red-team על ניתוח-Opus (gemini_local). **שער-קבע אוטומטי אחרי שלב הניתוח** — מופעל ע"י ה-CEO ברגע שהמנתח מסיים (`analysis-and-research.md`), **לפני** יצירת issue לכותב. הפלט = לידים לבדיקת-יו"ר בלבד (human-in-the-loop), ולעולם אינו מוזן לכותב אוטומטית. ראה למטה. |
 
-### שטן מליץ (Gemini) — דעה-שנייה on-demand בלבד ⚠️
+### שטן מליץ (Gemini) — שער-קבע אוטומטי אחרי הניתוח, עם עצירת-אישור ליו"ר ⚠️
 
-סוכן-Gemini שמבצע red-team על תוצר-המנתח (Opus) ומפיק **מזכר-לידים לא-סמכותי ליו"ר** (`critique-gemini.md`), read-only. **אינו נמצא בזרימת analyst→writer→qa.**
+סוכן-Gemini שמבצע red-team על תוצר-המנתח (Opus) ומפיק **מזכר-לידים לא-סמכותי ליו"ר** (`critique-gemini.md`), read-only. הוא **שער-קבע (standing gate) בזרימה**, שרץ **אוטומטית בין שלב הניתוח לבין הכתיבה** — אך הלידים שלו **נעצרים לאישור-יו"ר** ולעולם אינם זורמים לכותב אוטומטית.
 
-**מתי להפעיל:** **רק כשחיים/דפנה מבקשים מפורשות** "תן שטן-מליץ / דעה-שנייה על תיק X". אל תפעיל אותו אוטומטית, אל תכלול אותו בתזמור רגיל, ואל תציע אותו מיוזמתך.
+**מתי מופעל אוטומטית:** ברגע שהמנתח מסיים את הניתוח (`analysis-and-research.md` קיים; ראה "מפת סטטוסים" → `documents_ready`/`analyst_verified`) — אתה מפעיל את שטן-מליץ **לפני** שאתה ניגש לשלב B (סיכום + שאלת-תוצאה) ולפני יצירת issue כלשהו לכותב. אינך מחכה לבקשה מפורשת. (חיים עדיין יכול לבקש הרצה נוספת on-demand — אותו מנגנון בדיוק.)
 
-**כשמבקשים — איך:** צור issue המשויך ל-Agent ID של שטן-מליץ בחברה הנכונה (CMP=1xxx, CMPA=8xxx/9xxx) ו-wakeup רגיל עם `payload.issueId`.
+**איך מפעילים (זהה למסלול ה-on-demand הקודם):** צור issue המשויך ל-Agent ID של שטן-מליץ בחברה הנכונה (CMP=1xxx, CMPA=8xxx/9xxx) — עם `parentId` (ה-issue הראשי) וקישור `plugin_state` ל-case-number (ראה "כל issue חדש = תת-משימה"), ו-wakeup רגיל עם `payload.issueId`. ראה "שלב A2: שער שטן-מליץ" ו"שלב A3: עצירת-אישור הלידים" למטה לזרימה המדויקת.
 
-**הגבול הקריטי:** הפלט שלו = **לידים לבדיקת היו"ר בלבד** (human-in-the-loop). **אסור** להזין את הלידים שלו לכותב כמהות מאומתת, ואסור שיזרמו אוטומטית להחלטה. ה-writer ממשיך לצרוך **רק** את פלט-המנתח המעוגן. אם ליד של שטן-מליץ נראה חשוב — הוא עובר ליו"ר, היו"ר מאמת ומכריע, ורק אז (אם בכלל) הופך להנחיה.
+**הגבול הקריטי (לא נחלש — רק ה-trigger השתנה מ-on-demand ל-שער-קבע):** הפלט שלו = **לידים לבדיקת היו"ר בלבד** (human-in-the-loop). **אסור** להזין את הלידים שלו לכותב כמהות מאומתת, ואסור שיזרמו אוטומטית להחלטה. ה-writer ממשיך לצרוך **רק** את פלט-המנתח המעוגן + הנחיות-יו"ר מאושרות (chair_directions). אם ליד של שטן-מליץ נראה חשוב — הוא עובר ליו"ר, היו"ר מאמת ומכריע, ורק אז (אם בכלל) הופך ל-chair_direction דרך `approve_direction`. **השער הוא עצירה קשיחה לאישור-יו"ר — אסור להמשיך לכותב בלי אישור מפורש של היו"ר ללידים.** (מקיים G10 — שער אנושי; INV-AH/INV-LRN5 — לידים לא-סמכותיים, לא מוזנים אוטומטית.)
 
 ## כלל: כל issue חדש = תת-משימה
 
@@ -225,12 +226,15 @@ Paperclip חוסם אוטומטית כל issue ב-`in_progress` שאין לו ru
 ### שלב 0: בדוק למה התעוררת
 
 **לפני כל דבר אחר** — בדוק את סיבת ההתעוררות (`$PAPERCLIP_WAKE_REASON`):
-- אם ה-reason מכיל `user_commented` → **דלג ישירות לסעיף "טיפול בתגובות חדשות מחיים"**. אל תסרוק תיקים אחרים, אל תבדוק issues, אל תעשה heartbeat רגיל. **טפל רק בתגובה.**
+- **פעולות סטרוקטורליות (כפתורי-UI, בלי פענוח-טקסט):** אם `$PAPERCLIP_WAKE_PAYLOAD_JSON` מכיל שדה `action` → נתב דטרמיניסטית לפי הערך: `action == "interim_draft"` → **שלב H** (טיוטת ביניים); `action == "party_claims_summary"` → **שלב H2** (סיכום מנהלים). אלו side-quests — אל תסרוק תיקים אחרים, טפל רק בתיק שב-payload (`case_number`/`issueId`). ה-`issueId` שב-payload הוא **issue-ילד ייעודי שמשויך אליך** (נוצר ע"י המערכת כדי שהריצה לא תבוטל כשה-issue הראשי ממתין-ליו"ר; #227) — הרץ את הפעולה, פרסם comment עם התוצאה, ו**סגור את ה-issue-ילד הזה כ-done** בסיום. אל תיגע ב-issue הראשי של התיק.
+- אם ה-reason מכיל `user_commented` **ו**-`$PAPERCLIP_WAKE_PAYLOAD_JSON` כולל `issueId` → **דלג ישירות לסעיף "טיפול בתגובות חדשות מחיים"** עם ה-issue הזה. אל תסרוק תיקים אחרים. **טפל רק בתגובה.**
 - אם ה-reason מכיל `agent_completion` → דלג לשלב E/F בהתאם לסוכן שסיים
 - אם ה-reason מכיל `precedent_extraction_` → **דלג לסעיף "חילוץ פסיקה אוטומטי"**. אל תיגע בתיקים — זו עבודת ספרייה.
 - אם ה-reason מכיל `weekly-feedback-job` → **דלג לסעיף "ניתוח פידבק שבועי"**. אל תיגע בתיקים פעילים.
 - אם ה-reason מכיל `feedback_fold_` → **דלג לסעיף "קיפול הערת יו\"ר"**. אל תיגע בתיקים — זו משימת תחזוקת ידע.
-- אחרת → המשך לשלב A (heartbeat רגיל)
+- אחרת (כולל reason ריק) → ⚠️ **קודם הרץ את "סריקת הערות-יו\"ר תלויות" (שער-מסירה)**. רק אם לא נמצאה אף הערה תלויה — המשך לשלב A (heartbeat רגיל).
+
+> ⚠️ **wake עיוור — חובה לקרוא.** ה-sweep `route-pending-comments` (בפלאגין) מעיר אותך כשיש הערת-יו\"ר שטרם נענתה, דרך `ctx.agents.invoke` — אבל זה **אינו** מאכלס את `$PAPERCLIP_WAKE_REASON` ולא `payload.issueId` (אומת 2026-06-30, legal-ai #164: התעתיק הראה "generic heartbeat with no scoped wake"). לכן **הערת-יו\"ר תלויה מגיעה אליך כ-heartbeat "רגיל" בלי שום סימן** — ואם תסרוק רק את התור המשויך-אליך, היא תיפול. **לכן בכל heartbeat שאין בו reason מכוּון — חובה להריץ קודם את "סריקת הערות-יו\"ר תלויות".**
 
 ### חילוץ פסיקה אוטומטי
 
@@ -320,9 +324,9 @@ Paperclip חוסם אוטומטית כל issue ב-`in_progress` שאין לו ru
 ### שלב A: בדיקת מצב — שלמות, בדיקות שליליות, תאימות מתודולוגיה
 
 בכל heartbeat **רגיל** (לא comment routing):
-1. בדוק תיקים פעילים (`case_list`)
-2. בדוק אם יש issues ב-"blocked" — אם כן, טפל בהם קודם
-3. בדוק comments מחיים שממתינים לתגובה
+1. **קודם — "סריקת הערות-יו"ר תלויות"** (ראה סעיף ייעודי). ה-wake מה-sweep עיוור; הערת-יו"ר תלויה מגיעה כ-heartbeat רגיל. אם נמצאה הערה תלויה — טפל בה (scoped לתיק שלה) **לפני** כל דבר אחר.
+2. בדוק תיקים פעילים (`case_list`)
+3. בדוק אם יש issues ב-"blocked" — אם כן, טפל בהם קודם
 4. **לפני מעבר לשלב B — בצע את כל הבדיקות למטה. אם בדיקה נכשלת — עצור.**
 
 #### A1. בדיקת שלמות חילוץ
@@ -356,6 +360,57 @@ ls data/cases/$CASE_NUMBER/documents/research/analysis-and-research.md
 
 **עיקרון מנחה:** עדיף לעכב את התהליך מאשר לייצר החלטה על בסיס חלקי או פגום.
 
+> **⛔ סדר-זרימה מחייב אחרי שלב A — שער שטן-מליץ קודם לכותב.** ברגע ששלב A עבר (`analysis-and-research.md` מלא ותקין), **אסור** לקפוץ ישירות לשלב B/הכותב. הסדר הקשיח הוא: **מנתח סיים → שטן-מליץ (שלב A2, אוטומטי) → עצירת-אישור לידים ע"י היו"ר (שלב A3, in_review) → רק אחרי אישור-יו"ר, הלידים המאושרים → chair_directions → שלב B → … → כותב.** שטן-מליץ הוא שער-קבע, לא on-demand. ראה גם הערת "מפת סטטוסים" ל-`documents_ready`/`analyst_verified`.
+
+### שלב A2: שער שטן-מליץ (red-team אוטומטי אחרי הניתוח)
+
+**מתי:** שלב A עבר — `analysis-and-research.md` קיים ותקין, ועוד **לא** הופעל שטן-מליץ לתיק הזה (אין `critique-gemini.md`, ואין issue שטן-מליץ פתוח/סגור לתיק). זה השער הראשון אחרי הניתוח, **לפני** שלב B.
+
+1. **בדוק idempotency** — אם `data/cases/{case_number}/documents/research/critique-gemini.md` כבר קיים, או שכבר יצרת issue שטן-מליץ לתיק (חפש ב-issues הפעילים/הסגורים של החברה), **דלג** — השער כבר רץ. אם ה-critique קיים אבל עדיין לא הצגת אותו ליו"ר → לך ישר לשלב A3.
+2. **בחר את Agent ID של שטן-מליץ לפי חברה:** CMP (1xxx) = `9c86e06a-5a92-4723-af6d-e8cc6ae1d45b` · CMPA (8xxx/9xxx) = `46cc1228-a232-410b-a36b-71a6928499a2`.
+3. **צור issue** המשויך אליו (כותרת `[ערר {case_number}] שטן-מליץ — red-team על הניתוח`), עם `parentId=$PAPERCLIP_TASK_ID` ו-`description` שמפנה ל-`analysis-and-research.md` ומבקש להפיק `critique-gemini.md`. **חובה** את שני הצעדים מ"כל issue חדש = תת-משימה": (א) יצירת ה-issue, (ב) INSERT ל-`plugin_state` עם ה-case-number. ה-assignment מפעיל wakeup אוטומטי; אם נדרש wakeup ידני — `payload.issueId` של ה-issue החדש.
+4. **עדכן את ה-issue הראשי ל-`status=in_review`** (אתה ממתין לשטן-מליץ — אל תשאיר `in_progress` שייחסם). פרסם comment קצר: "הופעל שטן-מליץ (red-team) על הניתוח — אמתין למזכר-הלידים לפני המשך לשלב B."
+5. **אל תמשיך לשלב B כעת.** שטן-מליץ הוא read-only ואינו מעיר אותך בסיום (לפי `legal-analyst-gemini-critique.md` — הוא רק סוגר את ה-issue שלו). תתעורר עליו דרך heartbeat רגיל / סריקת-הערות; כשתראה ש-`critique-gemini.md` קיים → המשך לשלב A3.
+
+**הגבול הקריטי (חזרה):** הפלט של שטן-מליץ **אינו** נכנס לכותב ואינו הופך אוטומטית למהות. הוא קלט-יו"ר בלבד (G10 / INV-AH / INV-LRN5).
+
+### שלב A3: עצירת-אישור הלידים ע"י היו"ר (שער קשיח לפני הכתיבה)
+
+**מתי:** `critique-gemini.md` קיים (שטן-מליץ סיים), ועדיין לא הצגת את הלידים ליו"ר / לא קיבלת הכרעה.
+
+**זו עצירה מחייבת — אסור להמשיך לכותב בלי אישור-יו"ר ללידים.**
+
+1. **קרא** את `data/cases/{case_number}/documents/research/critique-gemini.md` במלואו.
+2. **פרסם comment ב-issue הראשי** עם **סיכום הלידים** — שמור על תיוג-הוודאות שהמזכר נשא (`[מאומת-קורפוס]`/`[טעון-אימות]`/`[ספקולציה]`), והפנֵה לקובץ:
+   ```
+   ## לידים של שטן-מליץ (red-team) — ערר {case_number}
+
+   המזכר המלא: `data/cases/{case_number}/documents/research/critique-gemini.md`
+   ⚠️ לידים לא-סמכותיים, טעוני-אימות יו"ר — לא הוזנו לכותב ולא להחלטה.
+
+   1. [תג-ודאות] {ליד מקוצר}
+   2. [תג-ודאות] {ליד מקוצר}
+   ...
+
+   אנא הכרע לכל ליד: אשר / דחה / הערה. רק לידים שתאשר יהפכו ל-chair_direction.
+   ```
+3. **צור interaction לבחירת לידים מאושרים** (`ask_user_questions`, `selectionMode: "multi"`, `idempotencyKey: "redteam-leads:{$PAPERCLIP_TASK_ID}:v1"`) — אופציה לכל ליד, plus היו"ר יכול להוסיף הערות ב-comment נפרד. (אם אין לידים מהותיים — interaction `request_confirmation` "אין לידים לאישור — להמשיך לשלב B?".)
+4. **עדכן את ה-issue הראשי ל-`status=in_review`** ושלח מייל:
+   ```bash
+   python3 /home/chaim/legal-ai/scripts/notify.py \
+     "נדרשת תשובתך — לידים של שטן-מליץ לתיק {case_number}" \
+     "שטן-מליץ הפיק מזכר-לידים על הניתוח. אנא בדוק ואשר/דחה לכל ליד לפני שנמשיך לכתיבה. קישור ל-issue."
+   ```
+5. **המתן.** אל תיגש לשלב B ואל תיצור issue לכותב — השער פתוח עד שהיו"ר מכריע.
+
+**קליטת הכרעת-היו"ר (כשמתעוררת עם `$PAPERCLIP_APPROVAL_ID` של interaction הלידים, או בתגובת-comment):**
+
+א. החזר את ה-issue הראשי ל-`status=in_progress`.
+ב. קרא את התשובה מה-API (`/interactions/$PAPERCLIP_APPROVAL_ID`) — אילו לידים אושרו; קרא גם comments אחרונים להערות-יו"ר.
+ג. **המר רק לידים מאושרים ל-chair_directions** — דרך **מנגנון ההנחיות הקיים** (אותו מנגנון של שלב 4 ב"טיפול בתגובות חדשות מחיים"): לכל ליד מאושר, קרא `record_chair_feedback(case_number, feedback_text="<הליד כפי שהיו\"ר אישר/חידד>", block_id="block-yod", category="missing_content"|"style"|"wrong_structure")`, **וגם** הוסף אותו ל-`analysis-and-research.md` תחת "עמדת ועדת הערר" בסוגיה המתאימה — כך `get_chair_directions(case_number)` יחזיר אותו לכותב, ובהמשך `approve_direction` יקפל אותו ל-direction_doc. **לידים שנדחו — מושלכים, לא נרשמים.**
+ד. פרסם comment: "הכרעת-יו"ר נקלטה. {N} לידים אושרו והומרו ל-chair_directions; {M} נדחו ונמחקו. ממשיך לשלב B." סגור את issue שטן-מליץ אם עוד פתוח.
+ה. **רק עכשיו** עבור לשלב B — עם הניתוח המעוגן + ההנחיות המאושרות בלבד. הכותב (בהמשך הזרימה) יצרוך **רק** את פלט-המנתח + chair_directions — **לעולם לא** את הלידים הגולמיים של שטן-מליץ.
+
 ### שלב B: הכנת סיכום, סיווג, ושאלת תוצאה
 
 **מתי:** כשיש `analysis-and-research.md` מלא (מנתח סיים שלבים 1-7) וסטטוס `analyst_verified`, אבל אין תוצאה עדיין
@@ -631,7 +686,9 @@ ls data/cases/$CASE_NUMBER/documents/research/analysis-and-research.md
 
 ### שלב H: טיוטת ביניים (לבקשת חיים, לפני דיון והכרעה)
 
-**מתי:** חיים מבקש בקומנט "טיוטת ביניים" / "interim draft" / "טיוטה לפני דיון" / "תכין לי את הטיוטה עם טענות הצדדים". בכל שלב לפני שיש תוצאה (בד"כ כשהתיק ב-`research_complete` או `analyst_verified`).
+**מתי — שתי דרכים:**
+- **(א) תגובת-יו"ר חופשית:** חיים מבקש בקומנט "טיוטת ביניים" / "interim draft" / "טיוטה לפני דיון" / "תכין לי את הטיוטה עם טענות הצדדים". בכל שלב לפני שיש תוצאה (בד"כ כשהתיק ב-`research_complete` או `analyst_verified`).
+- **(ב) פעולה סטרוקטורלית (כפתור-UI עתידי):** אם `$PAPERCLIP_WAKE_PAYLOAD_JSON` מכיל `action == "interim_draft"` (עם `case_number`/`issueId`) → הרץ את השלב הזה **דטרמיניסטית**, בלי פענוח-טקסט-חופשי. זה המסלול שכפתור "צור טיוטת ביניים" יפעיל. (אותה לוגיקה בדיוק כמו במסלול א — רק הטריגר שונה.)
 
 **מטרה:** ייצור מסמך עבודה לחיים עם פתיחה ניטרלית, רקע, תכניות+היתרים, טענות הצדדים, והליכים — **בלי דיון והכרעה**. חיים יכתוב את בלוק י בעצמו ואז נמשיך לזרימה הרגילה (QA + ייצוא סופי).
 
@@ -662,13 +719,13 @@ ls data/cases/$CASE_NUMBER/documents/research/analysis-and-research.md
    ```
    mcp__legal-ai__export_interim_draft(case_number="...")
    ```
-   מייצר `data/cases/{case_number}/exports/טיוטת-ביניים-v{N}.docx`, מעדכן `active_draft_path`.
+   מייצר `data/cases/{case_number}/exports/טיוטה-טענות_הצדדים_{N}.docx`, מעדכן `active_draft_path`.
 
 5. **דווח לחיים** (כולל מייל דרך `scripts/notify.py`):
    ```
    ## טיוטת ביניים מוכנה — ערר {case_number}
 
-   📄 **קובץ:** `data/cases/{case_number}/exports/טיוטת-ביניים-v{N}.docx`
+   📄 **קובץ:** `data/cases/{case_number}/exports/טיוטה-טענות_הצדדים_{N}.docx`
 
    ### מה כלול
    | בלוק | כותרת | מילים |
@@ -695,6 +752,49 @@ ls data/cases/$CASE_NUMBER/documents/research/analysis-and-research.md
 - אל תפעיל את שלב B (סיכום + שאלת תוצאה) במקביל — חיים מחליט מתי לעבור לזרימה הראשית.
 - אם בלוק ח חסר (אין פרוטוקול דיון/סיור) — ציין זאת בדוח. הכלי כותב מה שיש, אבל המשתמש צריך לדעת אם חסר.
 
+### שלב H2: סיכום מנהלים של טענות הצדדים (לבקשת חיים, לפני דיון)
+
+**מתי — שתי דרכים:**
+- **(א) תגובת-יו"ר חופשית:** חיים מבקש בקומנט "סיכום מנהלים" / "סיכום טענות" / "סיכום לקראת דיון" / "executive summary". בכל שלב שיש כבר טענות מאוגדות (`legal_arguments`) או claims גולמיים — בד"כ מ-`documents_ready`/`analyst_verified` ואילך.
+- **(ב) פעולה סטרוקטורלית (כפתור-UI עתידי):** אם `$PAPERCLIP_WAKE_PAYLOAD_JSON` מכיל `action == "party_claims_summary"` (עם `case_number`/`issueId`) → הרץ את השלב הזה **דטרמיניסטית**, בלי פענוח-טקסט-חופשי. זה המסלול שכפתור "צור סיכום מנהלים" יפעיל. (אותה לוגיקה בדיוק כמו במסלול א — רק הטריגר שונה.)
+
+**מטרה:** מסמך-פרוזה **מזוקק וקצר** של טענות הצדדים שמכין את היו"ר לדיון בעל-פה — תמצית-מנהלים, לא שכפול כתבי-הטענות. זהו מסמך **נפרד ומובחן** מטיוטת-ההחלטה **ומטיוטת-הביניים** (שלב H): סיכום-המנהלים אינו חלק מ-12 הבלוקים ואינו נכתב לתבנית ההחלטה — הוא נשמר ל-`documents/research/`. החלטת-יו"ר, #202.
+
+**זה side-quest, לא חלק מהזרימה B-F.** אל תשנה `cases.status`. אל תייצר issues לסוכני משנה. הכלי `summarize_party_claims` עושה הכל בעצמו (נגזר מ-`legal_arguments`/`claims` — אותו מקור-אמת, בלי חילוץ-מחדש; G2).
+
+**זרימה (~3-7 דקות):**
+
+1. פרסם comment קצר: "מתחיל יצירת סיכום מנהלים של טענות הצדדים — אעדכן בסיום." עדכן את ה-issue הראשי ל-`status=in_progress`.
+
+2. **הפקת הסיכום:**
+   ```
+   mcp__legal-ai__summarize_party_claims(case_number="...", instructions="")
+   ```
+   הכלי נגזר מ-`legal_arguments` (טיעונים מאוגדים) או, כ-fallback, מ-`claims` הגולמיים — אותו מקור-אמת של שלב הניתוח, בלי חילוץ-מחדש. הייצור עובר `claude_session` (`claude -p`) נעוץ ל-Opus 4.8 + `effort=high`, מתוחם לטענות-התיק בלבד (anti-hallucination). נשמר ל-`data/cases/{case_number}/documents/research/party-claims-summary.md`.
+
+3. **דווח לחיים** (כולל מייל דרך `scripts/notify.py`):
+   ```
+   ## סיכום מנהלים מוכן — ערר {case_number}
+
+   📄 **קובץ:** `data/cases/{case_number}/documents/research/party-claims-summary.md`
+
+   ### צפייה / הורדה / ייצוא
+   - קריאה (markdown): `GET /api/cases/{case_number}/research/party-claims-summary`
+   - הורדת .md: `GET /api/cases/{case_number}/research/party-claims-summary/download`
+   - ייצוא DOCX (סגנון-תבנית דפנה): `GET /api/cases/{case_number}/research/party-claims-summary/export-docx`
+
+   ### מה זה
+   תמצית-מנהלים מזוקקת של טענות הצדדים להכנה לדיון — **לא** טיוטת-החלטה ולא טיוטת-ביניים.
+   כולל פרק "נקודות-המחלוקת המרכזיות" כשאלות פתוחות, בלי הכרעה.
+   ```
+
+4. **סטטוס issue הראשי:** עדכן ל-`in_review` (חזרה להמתנה לחיים).
+
+**אזהרות:**
+- זהו מסמך **נפרד** מטיוטת-הביניים (שלב H). אל תבלבל ביניהם — אם חיים ביקש טיוטת-ביניים, רוץ שלב H; אם ביקש סיכום-מנהלים, רוץ שלב H2.
+- אל תייצא DOCX סופי (`export_docx`) ואל תפעיל את שלב B במקביל — זה side-quest.
+- אם אין טענות מאוגדות ולא claims (`summarize_party_claims` יחזיר שגיאה) — דווח לחיים שצריך קודם לסיים את חילוץ-הטענות/הניתוח, ואל תמציא תוכן.
+
 ## מפת סטטוסים
 
 **סטטוסים של התיק (`cases.status`) — כל סטטוס מתאים לפעולה אחת בדיוק:**
@@ -704,9 +804,9 @@ ls data/cases/$CASE_NUMBER/documents/research/analysis-and-research.md
 | `processing` | start-workflow (ממשק) | → בדוק אם כבר קיים issue פעיל לסוכן משנה. אם לא → המשך ל-§A כרגיל (בדוק documents + claims) |
 | `new` | (יצירת תיק) | → בדוק extraction_status של מסמכים. אם יש `pending` → צור issue למגיה (410c0167). אם כולם `completed`/`proofread` → צור issue למנתח |
 | `proofread` | מגיה | → צור issue למנתח משפטי (ראה תבנית למטה) |
-| `documents_ready` | מנתח | → שלב A (בדיקות שלמות + שליליות + מתודולוגיה). אם עובר → עדכן ל-`analyst_verified` |
-| `analyst_verified` | CEO (אחרי שלב A) | → שלב B (סיכום + שאלת תוצאה לחיים). המנתח כבר ביצע את המחקר כחלק מהניתוח — אין ליצור issue לחוקר. |
-| `research_complete` | מנתח / חוקר תקדימים (valid status — legacy + תרחישים מתקדמים) | → שלב B (סיכום + שאלת תוצאה לחיים). **זה סטטוס תקף**, לא שגיאה. בזרימה הרגילה המנתח מגדיר `documents_ready`, אבל אם החוקר רץ בנפרד (`legal-researcher.md` שלב 5) הוא מעדכן ל-`research_complete`. אם תראה סטטוס זה, בדוק שגם `analysis-and-research.md` וגם `precedent-research.md` קיימים, ואז המשך ל-§B כרגיל. |
+| `documents_ready` | מנתח | → שלב A (בדיקות שלמות + שליליות + מתודולוגיה). אם עובר → עדכן ל-`analyst_verified`, ואז **שלב A2 (שער שטן-מליץ, אוטומטי)** — אל תקפוץ לשלב B לפני שהשער רץ והיו"ר אישר לידים (A3) |
+| `analyst_verified` | CEO (אחרי שלב A) | → **קודם שער שטן-מליץ:** אם אין `critique-gemini.md` → שלב A2 (הפעל שטן-מליץ אוטומטית). אם יש `critique-gemini.md` שטרם הוצג ליו"ר → שלב A3 (עצירת-אישור לידים). **רק אחרי שהיו"ר אישר/דחה לידים** → שלב B (סיכום + שאלת תוצאה לחיים). המנתח כבר ביצע את המחקר כחלק מהניתוח — אין ליצור issue לחוקר. |
+| `research_complete` | מנתח / חוקר תקדימים (**סטטוס קנוני**, שלב "ניתוח וכיוון" — מקור-אמת: `case_status_model.py`) | → **קודם שער שטן-מליץ (A2/A3)** כמו ב-`analyst_verified`, ורק אחרי אישור-יו"ר ללידים → שלב B. **זה סטטוס תקף**, לא שגיאה. בזרימה הרגילה המנתח מגדיר `documents_ready`, אבל אם החוקר רץ בנפרד (`legal-researcher.md` שלב 5) הוא מעדכן ל-`research_complete`. אם תראה סטטוס זה, בדוק שגם `analysis-and-research.md` וגם `precedent-research.md` קיימים, ואז המשך לשער שטן-מליץ ואז §B. |
 | `outcome_set` | CEO (אחרי שחיים בחר) | → האם יש claim_handling? אם לא → שלב B המשך (טבלת bundle/skip). אם כן → שלב C |
 | `direction_approved` | CEO (אחרי שחיים אישר) | → צור issue למנתח (c26e9439) ל-pass 2: העמקת ניתוח ואימות פסיקה |
 | `analysis_enriched` | מנתח (pass 2) | → שלב D2: צור issue לכותב (7ed8686f) |
@@ -844,9 +944,49 @@ case_prefix="${case_number:0:1}"
 - **ודא עקביות מתודולוגית** — כיוונים סילוגיסטיים (כלל + עובדות + מסקנה), chair_directions שלם (טיפול בטענות + כיוון + סדר סוגיות + תקן ביקורת), התאמה ל-`decision-methodology.md`
 - **סינון תיקים** — עבוד רק על תיקים בטווח המספרים של החברה שלך (ראה טבלה למעלה)
 
+## סריקת הערות-יו"ר תלויות (company-wide) — שער-מסירה
+
+> **למה זה קיים:** ה-wake מה-sweep מגיע עיוור (ראה ⚠️ בשלב 0). לכן אינך יכול להסתמך על `reason`/`issueId` כדי לדעת שיש הערה לטפל בה — אתה חייב **לגלות** אותה בעצמך. סריקה זו מוצאת כל הערת-יו"ר שטרם נענתה **בכל issue בחברה** (לא רק בתור המשויך-אליך — הערות-יו"ר יושבות על issues של חיים/סוכני-מִשנה, לא עליך).
+
+**⚠️ כלל-ברזל: שיוך-לתיק ממקור מובנה, לעולם לא מהכותרת ולא מתוכן-ההערה.** הערה עשויה להזכיר תיק אחר (למשל "ניסיתי ב-CMPA-89") — זה **תוכן לטיפול**, לא יעד. היעד הוא תמיד התיק של ה-**project** שאליו שייך ה-issue. אם אי-אפשר לפתור תיק בוודאות → **אל תפעל**.
+
+**הסריקה (הרץ בכל heartbeat ללא reason מכוּון):**
+
+1. **רשום את כל issues החברה הפעילים:**
+   ```bash
+   ISSUES=$(~/legal-ai/scripts/pc.sh GET "/api/companies/$PAPERCLIP_COMPANY_ID/issues")
+   echo "$ISSUES" | jq -r '(if type=="array" then . else .issues end)
+     | map(select(.status!="done" and .status!="cancelled" and (.hiddenAt|not)))
+     | .[].id'
+   ```
+
+2. **לכל issue — האם יש הערת-יו"ר תלויה?** הערה תלויה = הערת-`user` האחרונה **חדשה** מהערת-`agent` האחרונה (התעלם מ-`system`). ⚠️ ה-API מחזיר comments ב-**newest-first** — אל תשתמש ב-`.[-1]`; מיין ב-`createdAt`:
+   ```bash
+   C=$(~/legal-ai/scripts/pc.sh GET "/api/issues/$ISSUE_ID/comments")
+   PENDING=$(echo "$C" | jq -r '
+     (map(select(.authorType=="user"))  | sort_by(.createdAt) | last) as $u |
+     (map(select(.authorType=="agent")) | sort_by(.createdAt) | last) as $a |
+     if $u==null then "no" elif ($a==null) or ($u.createdAt>$a.createdAt) then "yes" else "no" end')
+   ```
+   אם `PENDING != "yes"` → דלג ל-issue הבא.
+
+3. **שיוך-לתיק (שער-בטיחות חובה)** — חלץ את מספר-התיק מ-**`project.name`** (שדה מובנה: `"ערר <מספר> — ..."`), לא מהכותרת ולא מההערה:
+   ```bash
+   CTX=$(~/legal-ai/scripts/pc.sh GET "/api/issues/$ISSUE_ID/heartbeat-context")
+   CASE_NUMBER=$(echo "$CTX" | jq -r '.project.name // ""' \
+     | grep -oE '(ערר|בל"מ) [0-9]{4,5}-[0-9]{2}-[0-9]{2}' \
+     | grep -oE '[0-9]{4,5}-[0-9]{2}-[0-9]{2}' | head -1)
+   ```
+   - **`CASE_NUMBER` ריק** (ה-project אינו תיק — למשל "ספריית פסיקה — תור חילוץ", או שיוך לא-ודאי) → **אל תפעל על תיק.** אם נראה שזו הערת-יו"ר אמיתית שאבדה — פרסם comment קצר ב-issue + שלח מייל ליו"ר (סעיף §5), בלי לגעת בשום תיק. עבור ל-issue הבא.
+   - **`CASE_NUMBER` קיים** → טפל בהערה לפי "טיפול בתגובות חדשות מחיים" למטה, **scoped ל-`$ISSUE_ID` ול-`$CASE_NUMBER` הזה בלבד.**
+
+4. **כמה issues תלויים** (אפשרי — אפילו על אותו תיק) → טפל בכולם, כל אחד ב-scope שלו. אם לא נמצאה אף הערה תלויה → אין פעולת-CEO; המשך לשלב A.
+
+---
+
 ## טיפול בתגובות חדשות מחיים (comment routing)
 
-כשאתה מתעורר בגלל תגובה חדשה (reason מכיל "user_commented"):
+כשאתה מתעורר בגלל תגובה חדשה (reason מכיל "user_commented"), **או** כשהסריקה למעלה זיהתה הערה תלויה ב-`$ISSUE_ID` עם `$CASE_NUMBER` פתור:
 
 0. **החזר את ה-issue הראשי ל-`status=in_progress`** — אם ה-issue ב-`in_review` (כי המתנת לחיים) או ב-`blocked` (כי Paperclip חסם אוטומטית), הראשון דבר: עדכן ל-`in_progress` כדי לסמן שאתה עובד עליו.
 

.claude/agents/legal-qa.md

@@ -179,10 +179,11 @@ tools:
 בדוק `case_get(case_number).status` — הוא צריך להיות בערכים תקפים. הזרימה הכוללת:
 
 ```
-new → proofread → documents_ready → analyst_verified → research_complete (legacy/optional)
-   → outcome_set → direction_approved → analysis_enriched → ready_for_writing
-   → drafted (אתה כאן!) → qa_passed / qa_failed → exported
+new → processing → documents_ready → analyst_verified → research_complete
+   → outcome_set → direction_approved → qa_review → drafted (אתה כאן!)
+   → exported → reviewed → final
 ```
+(מקור-אמת יחיד: `case_status_model.py` / `GET /api/status-model`. `analyst_verified`+`research_complete` קנוניים, שלב "ניתוח וכיוון". מצבי-ביניים ישנים כמו `proofread`/`analysis_enriched`/`qa_passed` נמפים-לשלב לתצוגה בלבד.)
 
 ⚠️ **`research_complete` הוא valid status** (לא bug, לא legacy ערומה). ב-`legal-researcher.md` שלב 5 הוא הסטטוס שהחוקר מגדיר בסיום מחקר. אם תיק במצב זה נשלח אליך לפני `drafted` — דווח, אל תכשיל.
 

.gitignore

@@ -19,6 +19,9 @@ kiryat-yearim/
 continuation-prompt.md
 node_modules/
 data/eval/eval-report-*
+# effort-calibration run outputs (#208 / #212) — generated, not source
+data/eval/effort-calibration-*
+data/eval/block-yod-calibration-*
 # revert snapshot for migrate_agent_adapter.py (runtime state)
 data/adapter-migration-state.json
 # frontmatter-stripped instruction copies for content_arg adapters (generated)

docs/paperclip-quirks.md

@@ -275,3 +275,40 @@ WARN [plugin] issue.comment.created event missing issueId in payload, skipping
 - **תוקן בצד שלנו** (PR #2, נפרס דרך `npm run build` + `pm2 restart paperclip` — הפלאגין נטען מ-`/home/chaim/plugin-legal-ai` לפי `package_path`).
 - **לקח כללי**: ל-events של הפלאגין — מזהה-הישות-הראשית הוא תמיד `event.entityId`; אל תניח ששדות נמצאים ב-`payload` בלי לאמת מול ה-`.d.ts` של ה-SDK או מול לוג חי.
 - TaskMaster: `legal-ai` #149.
+
+---
+
+## 7. `agents.invoke({prompt})` — הפרומפט נבלע; רק `payload.issueId` נמסר
+
+**התסמין (מנקודת-מבט היו"ר):** כותבים הוראה לסוכן — והוא "מתעלם". ההרצה נראית תקינה
+לחלוטין: `status=succeeded`, `exit_code=0`, בלי שגיאה, ואפילו עולה כסף. הסוכן מדווח
+"כל התורים ריקים — אין עבודה".
+
+**הסיבה:** `ctx.agents.invoke(agentId, companyId, {prompt, reason})` הוא ה-API **המתועד**
+של ה-Plugin SDK ("*Invoke an agent with a prompt payload*"), ו-`prompt` הוא שדה חובה.
+Paperclip מקבל אותו, שומר אותו ב-`agent_wakeup_requests.payload.prompt` — **וה-runner קורא
+רק `payload.issueId`**. הפרומפט לא מגיע למודל. ה-`reason` כן שורד (ל-`context_snapshot.wakeReason`),
+ולכן זה נראה כאילו משהו נמסר.
+
+מדידה ב-DB (2026-07-16): **18 מתוך 18** ה-wakeups שאי-פעם נשאו `payload.prompt` לא הגיעו
+לאף הרצה — 0% מסירה. מול 107,172 wakeups עם `issueId` שעובדים. אימות ישיר: בלוג-ההרצה
+שנבלעה אין ולו אזכור אחד של מספר-התיק, והמודל פותח ב-*"This heartbeat has no scoped wake payload"*.
+
+**מה עושים במקום:** נושאים את ההוראה על **issue**, לא ב-payload — issue-ילד שכבר משויך
+ל-CEO, עם ההוראה ב-`description`, ואז wakeup עם `payload.issueId`. זה בדיוק
+`open_ceo_run` ב-[`web/paperclip_client.py`](../web/paperclip_client.py), והמסלול היחיד
+שאומת כמגיע לסוכן.
+
+**מלכודות שנלוות לזה — אל תיפול בהן שוב:**
+- **`issues.create` של ה-SDK עם `assigneeAgentId` *לא* מעיר את הסוכן.** ה-REST המקביל
+  (`POST /api/issues/:id/children`) *כן* — הוא פולט wakeup `issue_assigned`. אל תכליל
+  מהתנהגות REST על ה-SDK; זה נבדק ונמצא שונה.
+- **`ctx.issues.requestWakeup` (ה-primitive הנכון ב-SDK) נכשל מתוך scheduled job** עם
+  `missing, expired, or unknown invocation scope` — למרות ש-capability `issues.wakeup`
+  מוצהר ומותקן. אותה שגיאה מפילה גם `companies.list` / `issues.list` / `issues.listComments`
+  בתוך ה-sweep, ולכן **רשת-הביטחון `route-pending-comments` אינה אמינה**. לא נחקר לעומק —
+  לכן ניתוב-ההערות הועבר ל-backend של legal-ai (REST), שאינו חשוף לבעיה.
+
+**סטטוס:** נעקף — `POST /api/cases/{case}/agents/comment` ו-`/agents/interaction-response`
+פותחים הרצת-CEO ישירות (REST) ומסמנים את ההערה כמנותבת, כדי שה-sweep השבור לא יירה
+הרצת-סרק. TaskMaster `legal-ai` #228.

docs/spec/01-ingest.md

@@ -96,10 +96,11 @@ idempotency-key pattern · ISO 8000 (Data quality) | סטטוס: verified
 בנקודת-סיום-הקליטה, לכל סוג; חוזה-שלמות יסמן רשומה ללא מטא-דאטה כלא-שמישה
 ([G4](00-constitution.md#inv-g4-חוזה-שלמות-לפני-שמיש--ניתן-לחיפוש), מפורט ב-
 [02-data-model.md](02-data-model.md)).
-**הפרה ידועה:** המסלול הפנימי (`internal_decisions.py:208`) מתזמן **רק**
-`request_halacha_extraction` ואינו קורא ל-`request_metadata_extraction` (בניגוד
-ל-`precedent_library.py:292-293` שקורא לשניהם) → ערן סופר 8046/24 נקלטה **בלי
-metadata** (headnote/summary/tags ריקים) → ממצא ל-[audit](../audit-report.md).
+**הפרה ידועה — ✅ נסגרה (FU-1, GAP-02):** המסלול הפנימי דילג בעבר על חילוץ-המטא-דאטה
+(`internal_decisions.py:208` תיזמן רק `request_halacha_extraction`) → ערן סופר 8046/24 נקלטה
+בלי metadata. **נסגר:** שני מסלולי-הפסיקה זורמים כעת דרך `ingest.ingest_document` הקנוני, שמתזמן
+**את שני התורים יחד** בנקודה אחת (`ingest.py:233-234`) — תיזמון-המטא אינו מותנה במסלול עוד.
+מאומת ב-`test_unified_ingest.py` (9/9). תיעוד-מקור: [audit](../audit-report.md).
 
 ### INV-ING4: re-index בקליטה-חוזרת (upsert ⇒ re-embed)
 **כלל:** קליטה-חוזרת ששינתה את תוכן-הפריט מפעילה **re-index** — chunks ו-embeddings

docs/spec/02-data-model.md

@@ -73,6 +73,9 @@ proceeding_type)`. לכן המזהה הקנוני הוא **(`case_number` מנו
 ### 2ב. חוזה תמציתי לישויות נוספות
 
 - `documents` → usable: `file_path`+`doc_type`; searchable: `extraction_status=completed` ו-`extracted_text` לא-ריק ו-≥1 `document_chunk` עם embedding.
+  > **עיצוב-מחדש זרימת-העבודה (WS2, #200):** `documents` מקבל `is_primary`/`doc_category` (רשימה
+  > קנונית: ערר/תשובה/תגובה/פרוטוקול/שומה/החלטת-ועדה) + דגל-ניתוח per-מסמך (`claims_extraction_status`),
+  > כדי לחשב "יש מסמך-עיקרי שלא-נכלל-בניתוח". השדות בבעלות #200; ניתוח-החילוץ עצמו ב-[04-analysis-writing](04-analysis-writing.md) §1.
 - `halachot` → usable: `rule_statement`+`supporting_quote`; **searchable: `review_status ∈ {approved, published}` בלבד** — `pending_review`/`rejected` מוסתרות מ-`search_precedent_library` (שער-הלכה ידני, `db.py:644-660`, [G10](00-constitution.md#inv-g10-המערכת-מסייעת--שערים-אנושיים-הם-invariant)).
 - `decision_blocks` → usable: `block_id`∈12-הבלוקים; "מוכן": `status=final` ו-`content` לא-ריק.
 - `chair_feedback` → usable: `feedback_text`+`category` מהמילון; "פתוח" עד `resolved=true`.
@@ -207,18 +210,22 @@ RAG freshness (Lewis et al., 2020, NeurIPS) | סטטוס: verified
   (`case_number, source_kind, proceeding_type`, `db.py:904-909`), אך השדה `case_number TEXT
   UNIQUE NOT NULL` המקורי (`db.py:368`) הוסר רק ב-V15 (`db.py:902-903`) — מורשת שאפשרה את
   הפרת [INV-DM2](#inv-dm2-מזהה-קנוני-יחיד-לכל-ישות). **יעד:** נרמול-בכתיבה אכוף + ציטוט-מלא רק ב-`citation_formatted`.
-- **`summary` קיים על `case_law` אך לא בחוזה-הקליטה הפנימי.** העמודה קיימת (`db.py:373`) אך
-  המסלול הפנימי אינו ממלא אותה (כפועל-יוצא מהיעדר חילוץ-מטא-דאטה, [INV-ING3](01-ingest.md#inv-ing3-תור-חילוץ-מטא-דאטה--הלכות-לכל-סוג)).
-  **יעד:** searchable מותנה ב-metadata לא-ריק ([INV-DM1](#inv-dm1-searchable-רק-כשחוזה-השלמות-מתקיים)).
-- **שני שדות-סטטוס-חילוץ נפרדים, ללא דגל-`searchable` מפורש.** `extraction_status` +
-  `halacha_extraction_status` (`db.py:603-605`) מתארים תהליך, אך אין שדה יחיד שמסמן "עבר
-  חוזה-שלמות → searchable". **יעד:** דגל/view נגזר ש-search מסנן לפיו, מגובה health-check.
+- **`summary` קיים על `case_law` אך לא בחוזה-הקליטה הפנימי — ✅ נסגר (FU-1, GAP-02).** העמודה
+  קיימת (`db.py:373`), והמסלול הפנימי לא מילא אותה בעבר (היעדר חילוץ-מטא-דאטה). **נסגר:** שני
+  המסלולים זורמים דרך `ingest.ingest_document` הקנוני שמתזמן metadata לשני הסוגים
+  ([01-ingest INV-ING3](01-ingest.md#inv-ing3-תור-חילוץ-מטא-דאטה--הלכות-לכל-סוג)); `searchable`
+  מותנה ב-metadata לא-ריק ([INV-DM1](#inv-dm1-searchable-רק-כשחוזה-השלמות-מתקיים)).
+- **דגל-`searchable` מפורש — ✅ קיים.** מעבר ל-`extraction_status` + `halacha_extraction_status`
+  (`db.py:603-605`, תהליך), קיימת כעת עמודת `case_law.searchable` המסמנת "עבר חוזה-שלמות"; שכבת-
+  החיפוש מסננת לפיה (`cl.searchable = true`, `db.py:7517,7519`) ו-`recompute_searchable_flag`
+  מתחזק אותה. אוכף את [INV-DM1](#inv-dm1-searchable-רק-כשחוזה-השלמות-מתקיים) בשכבת-האחזור.
 - **`embedding` אינו `GENERATED` (בניגוד ל-tsvector).** ה-tsvectors מסונכרנים אוטומטית
   (`db.py:776,780,1083`), אך ה-`embedding vector(1024)` תלוי-טריגר חיצוני — נקודת-drift אפשרית
   ל-[INV-DM3](#inv-dm3-שינוי-תוכן--re-index). **יעד:** טריגר re-embed מובטח + health-check ל-drift.
-- **`halachot.review_status` כשער-searchable ללא נראות-backlog.** הסינון תקין (`pending_review`
-  מוסתר, `db.py:659`), אך אין נראות כמה ממתינות — תואם את ההפרה הידועה ב-[G10](00-constitution.md#inv-g10-המערכת-מסייעת--שערים-אנושיים-הם-invariant)
-  (10/19 מאושרות, התגלה במקרה). **יעד:** health-check חושף backlog-הלכות.
+- **נראות backlog-הלכות — ✅ נסגר (FU-5, GAP-14).** הסינון תקין (`pending_review` מוסתר,
+  `db.py:659`); ההפרה הישנה היתה היעדר-נראות (10/19 מאושרות, התגלה במקרה). **נסגר:** ה-health-
+  check חושף `halacha_backlog` (`web/app.py:2352-2364`, `metrics_service.halacha_backlog`),
+  והכלים `extraction_status` + `halachot_pending` חושפים את עומק-התור — שער-היו"ר ([G10](00-constitution.md#inv-g10-המערכת-מסייעת--שערים-אנושיים-הם-invariant)) הופך נראה.
 
 ---
 

docs/spec/03-retrieval.md

@@ -79,11 +79,12 @@ quality / conformance) | סטטוס: verified
 **אכיפה:** תנאי-`source_kind` בכל ענף-SQL בשכבת-החיפוש; בדיקת-בריאות שמריצה שאילתת-ביקורת
 (חיפוש מכוון-קורפוס שמחזיר פריט בעל `source_kind` זר = כשל). אוכף את
 [G5](00-constitution.md#inv-g5-metadata-מלא--הפרדת-קורפוס-נאכפת-בכל-query).
-**הפרה ידועה:** משימה #56 — `halacha_filters` **אינם** כוללים `cl.source_kind` ב-
-`search_precedent_library_semantic` (`db.py:3168`, ענף ה-halacha; לעומת `chunk_filters` שכן —
-`db.py:3169`) **וב**-`search_precedent_library_lexical` (`db.py:3401` מול `db.py:3402`). שני
-ה-`halacha_sql` עושים `JOIN case_law cl` בלי לסנן `source_kind` (`db.py:3236-3238`, `db.py:3475-3477`)
-→ הלכות מהקורפוס הפנימי דולפות לתוצאות החיפוש בקורפוס החיצוני ולהפך → ממצא ל-[audit](../audit-report.md).
+**הפרה ידועה — ✅ נסגרה (FU-4, GAP-10, משימה #56):** בעבר `halacha_filters` לא כללו
+`cl.source_kind` (בניגוד ל-`chunk_filters` שכן) → הלכות מהקורפוס הפנימי דלפו לחיפוש בקורפוס
+החיצוני ולהפך. **נסגר:** שני ה-`halacha_filters` (semantic + lexical) מתחילים כעת ב-
+`cl.source_kind = '{source_kind}'` בדיוק כמו `chunk_filters` (`db.py:7516,7519` ומקבילו ב-
+`search_precedent_library_lexical`). מאומת ב-`test_precedent_corpus_isolation.py` (semantic+lexical,
+שני ה-source_kind) → תיעוד-מקור: [audit](../audit-report.md).
 
 ### INV-RET2: אין החזרה/אינדוקס בלי metadata מלא + locator פתיר
 **כלל:** פריט אינו מוחזר מ-search (ואינו נחשף לאחזור) אלא אם **שדות-החובה שלו מולאו**
@@ -121,10 +122,13 @@ relevance judgments, precision/recall) · RAG evaluation literature (Lewis et al
 Elastic — *relevance evaluation guidance* | סטטוס: verified
 **אכיפה:** eval harness עם gold-set מתויג + בדיקת-בריאות תקופתית; שער-CI על שינוי שכבת-האחזור.
 אוכף את [G8](00-constitution.md#inv-g8-איכות-אחזור-נמדדת--precision--recall).
-**הפרה ידועה (GAP):** אין כיום eval harness ולא gold-set — קיים רק `telemetry.log_search_bg`
-(`search.py:62,118,190,271`; `precedent_library.py:280`) שמתעד שאילתות בפועל, אך **אינו מודד
-precision/recall מול תיוג** (תצפית, לא הערכה). היעד: harness שמריץ סט קבוע ומחזיר metrics →
-ממצא ל-[audit](../audit-report.md).
+**הפרה ידועה — ✅ נסגרה (FU-5, GAP-11):** בעבר היה רק `telemetry.log_search_bg` (תצפית, לא
+הערכה). **נסגר:** `scripts/eval_retrieval.py` מריץ את **מסלול-האחזור בייצור** (אותן פונקציות-
+שירות שכלי-ה-MCP קוראים) מול gold-set מתויג (`data/eval/gold-set.jsonl`, נבנה ב-
+`scripts/eval_gold_bootstrap.py`) ומחזיר precision/recall/MRR/nDCG@{5,10} — סך-הכול, פר-קורפוס
+ופר-`practice_area` — מול baseline מקובע (`data/eval/baseline.json`). מטריקות-הליבה נבדקות
+offline ב-`eval_retrieval.py --self-test` ("ALL PASS"). **משמעת-CI:** להריץ לפני/אחרי כל שינוי
+שכבת-אחזור (משקלי-RRF/`k`/סף-chunk/embedder/rerank) ולהשוות ל-baseline. תיעוד-מקור: [audit](../audit-report.md).
 
 ### INV-RET5: כל span מוחזר עקיב למקורו
 **כלל:** כל קטע מוחזר נושא **עקיבוּת-מקור מלאה** — מזהה-מסמך/פסק-דין (`case_law_id`/`document_id`/
@@ -156,16 +160,15 @@ re-embed; בדיקת-בריאות מגלה embeddings מיושנים. אוכף
 
 ההבדלים בין הקוד בפועל ל-TARGET. **אלו תסמינים, לא התנהגויות תקינות.** כל פריט אומת מול הקוד.
 
-- **דליפת-הלכות חוצת-קורפוס (משימה #56).** `halacha_filters` נפתחים רק עם `review_status`
-  (`db.py:3168`, `db.py:3401`) ואינם מוסיפים `cl.source_kind`, בעוד `chunk_filters` כן
-  (`db.py:3169`, `db.py:3402`). שני ה-`halacha_sql` עושים `JOIN case_law` בלי סינון
-  (`db.py:3236-3242`, `db.py:3463-3482`). **תסמין:** חיפוש בקורפוס החיצוני
-  (`search_precedent_library`, `source_kind="external_upload"`) יכול להחזיר הלכה שמקורה
-  בהחלטת-ועדה פנימית — ולהפך עבור `search_internal_decisions` (`source_kind="internal_committee"`,
-  `internal_decisions.py:418`). **יעד:** `halacha_filters` יתחילו ב-`cl.source_kind = '{source_kind}'`
-  בדיוק כמו `chunk_filters` ([INV-RET1](#inv-ret1-הפרדת-קורפוס-נאכפת-ב-100-ממסלולי-ה-query)).
-- **אין eval harness — מדידת-איכות לא קיימת.** רק `telemetry.log_search_bg` מתעד שאילתות
-  (`search.py:62,118,190,271`); אין gold-set מתויג ואין precision/recall. **יעד:** harness עומד
+- **דליפת-הלכות חוצת-קורפוס (משימה #56) — ✅ נסגר (FU-4, GAP-10).** בעבר `halacha_filters` לא
+  כללו `cl.source_kind` (בניגוד ל-`chunk_filters`) → הלכה מהקורפוס הפנימי יכלה לחזור בחיפוש
+  בקורפוס החיצוני, ולהפך. **נסגר:** `halacha_filters` מתחילים כעת ב-`cl.source_kind = '{source_kind}'`
+  + `cl.searchable = true` בדיוק כמו `chunk_filters` (`db.py:7514-7519`), בשתי הפונקציות
+  (semantic+lexical). מאומת ב-`test_precedent_corpus_isolation.py`
+  ([INV-RET1](#inv-ret1-הפרדת-קורפוס-נאכפת-ב-100-ממסלולי-ה-query)).
+- **eval harness — ✅ נסגר (FU-5, GAP-11).** בעבר רק `telemetry.log_search_bg` (תצפית).
+  **נסגר:** `scripts/eval_retrieval.py` (precision/recall/MRR/nDCG מול `data/eval/gold-set.jsonl`
+  + baseline; `--self-test` עובר offline) — מדידה אמפירית מול תיוג קבוע
   ([INV-RET4](#inv-ret4-איכות-אחזור-נמדדת-ב-eval-harness-עומד-precision--recall)).
 - **`search_decisions` מתעד אזהרה כשאין `practice_area` אך לא חוסם.** ללא פילטר-תחום החיפוש
   עלול לערבב תחומים משפטיים (`search.py:45-49,172-176` — `logger.warning`, ממשיך). **יעד:** הפרדה

docs/spec/04-analysis-writing.md

@@ -47,6 +47,17 @@
 > מקוריות מהשלמות). הסינון לפי `party_role` מאפשר לזהות את הצד המפסיד ולוודא שכל טיעון
 > שלו מקבל מענה בבלוק י.
 
+> **עיצוב-מחדש זרימת-העבודה (WS2–WS4) — מסלול-החילוץ אינו מתפצל ([G2](00-constitution.md#inv-g2-מקור-אמת-יחיד--אין-מסלולים-מקבילים-מתפצלים)):**
+> ניתוח-ערר-לבד כבר עובד (`extract_claims` רץ על `appeal` בנפרד). ההרחבות נשענות על אותו צינור-
+> חילוץ, **לא** על מסלול-ניתוח שני: (א) **ניתוח-מחדש מאחד** (WS2, #200/#201) — מחלץ מ-מסמכים
+> חדשים/מעודכנים בלבד ומאחד עם הקיים (לא מוחק-ומחשב-מחדש), על-בסיס דגל-הניתוח per-מסמך
+> ([02-data-model §2ב](02-data-model.md#2ב-חוזה-תמציתי-לישויות-נוספות)); (ב) **סיכום-מנהלים**
+> (WS3, #202, `summarize_party_claims` — ראה §1.4) — נגזר מ-`claims`/`legal_arguments`, **מסמך-פרוזה נפרד**
+> מטיוטת-הביניים (G2: תצוגה-נגזרת, לא מקור-אמת שני); (ג) **ניתוח-פרוטוקול** (WS4, #203,
+> `analyze_protocol` — ראה §1.3) — `protocol` נכנס לחילוץ ההשוואתי (ירד/חוזק/חדש) ומזין ידע-תיק, ונשאר
+> פוסט-דיון ([INV-WR4](#inv-wr4-בלוק-ז--טענות-מקוריות-בלבד): טענות-פוסט-דיון → בלוק ח, לא ז).
+> חוזי-הכלים המלאים בבעלות המשימות הנ"ל (§1.3/§1.4); כאן רק עיגון-הספ שהם מקיימים G2.
+
 ### 1.3 ניתוח-פרוטוקול — ניתוח פרוטוקול-דיון השוואתי (WS4 / #203)
 
 > **בעלוּת תת-סעיף:** §1.3 (זה) שייך ל-WS4 (ניתוח-פרוטוקול). תת-סעיף סיכום-טענות-הצדדים
@@ -74,6 +85,11 @@ knowledge נגזר; G2: מקור-האמת הוא הפרוטוקול + `legal_arg
   היו"ר כבר מילא תאריך — לא דורסים קלט-יו"ר).
 - **ייצור:** קריאת-ה-LLM ההשוואתית עוברת `claude_session` (מקומי בלבד), מעוגנת
   `model="claude-opus-4-8"` + `effort="high"` (ראה `reference_claude_generation_path`).
+  **חלון-הקשר (#216):** הבנייה הסטנדרטית רצה 200K-טוקן (~400K תווים עברית). פרומפט גדול
+  (בעיקר בלוק-י, שנושא את התיק המלא כ-source-context) מוסלם אוטומטית לבניית-ה-1M של *אותו*
+  מודל — `claude-opus-4-8[1m]` (1M-טוקן, מחיר רגיל; `claude -p` חושף אותה ישירות כ-model-id,
+  בלי beta-header). **עידון של אותו pin, לא מסלול-מודל מקביל (G2).** הסף ב-`block_writer`:
+  פרומפט > ~350K תווים → `[1m]`; אחרת הבנייה הסטנדרטית. תקרת-קשיחה: 1.5M תווים.
 
 ### 1.4 סיכום-מנהלים של טענות הצדדים (מסמך-הכנה לדיון, WS3/#202)
 
@@ -96,6 +112,12 @@ knowledge נגזר; G2: מקור-האמת הוא הפרוטוקול + `legal_arg
   (git + S3, באותו מסלול-אחסון של `analysis-and-research.md`); ניתן-לייצוא ל-DOCX בסגנון-תבנית
   דפנה (`build_party_claims_summary_docx`). מימוש: `tools/drafting.py` ·
   `services/party_claims_summary.py` · `services/analysis_docx_exporter.py`.
+- **טריגר — side-quest מבוקש-יו"ר (G10), נפרד מטיוטת-הביניים:** ה-CEO מפיק את הסיכום כשלב-צד
+  (`legal-ceo.md` שלב H2) ב-**שתי דרכים**: (א) הערת-יו"ר חופשית ("סיכום מנהלים" / "סיכום טענות" /
+  "סיכום לקראת דיון" / "executive summary"); או (ב) **פעולה סטרוקטורלית** —
+  `$PAPERCLIP_WAKE_PAYLOAD_JSON` עם `action == "party_claims_summary"` (המסלול הדטרמיניסטי שכפתור-UI
+  עתידי יפעיל, בלי פענוח-טקסט). זהו side-quest: אינו משנה `cases.status` ואינו יוצר issues לסוכני-משנה,
+  ומובחן מטיוטת-הביניים (`write_interim_draft`, שלב H, `action == "interim_draft"`).
 
 ### 1.5 ניתוח-מחדש מאחד אחרי מסמך-עיקרי חדש (#201)
 
@@ -117,6 +139,30 @@ reanalyze_all_primary=False)` סוגר את הפער **בלי force-delete גו
 
 ראה `tools/legal_arguments.py` (`reanalyze_claims`, `_snapshot`, `_impact_diff`).
 
+### 1.6 שער שטן-מליץ (red-team) — בין הניתוח לכתיבה, תחת אישור-יו"ר
+
+אחרי שלב-הניתוח (`analysis-and-research.md` תקין) וב**לפני** הכותב, ה-CEO מפעיל **אוטומטית**
+שכבת **דעה-שנייה אדוורסרית** מ-lineage שונה (Gemini, `legal-analyst-gemini-critique`,
+read-only) — **שער-קבע (standing gate)**, לא on-demand. השכבה תוקפת את ניתוח-Opus ומפיקה
+`critique-gemini.md` = **מזכר-לידים לא-סמכותי**, מתויג-ודאות (`[מאומת-קורפוס]`/`[טעון-אימות]`/
+`[ספקולציה]`), כפוף לשער ה-anti-hallucination ([INV-AH](../anti-hallucination-gate.md);
+כלי-RAG משפטיים הוזים פסיקה 17–33%, Stanford RegLab/Magesh JELS 2025).
+
+- **human-in-the-loop קשיח ([G10](00-constitution.md#inv-g10-המערכת-מסייעת--שערים-אנושיים-הם-invariant)):**
+  ה-CEO מציג את הלידים ליו"ר כ**עצירת-אישור** (ה-issue הראשי ל-`in_review` + מייל), ואינו מתקדם
+  לכותב בלי הכרעת-יו"ר מפורשת. **רק לידים שהיו"ר אישר** מומרים ל-`chair_directions` דרך
+  מנגנון-ההנחיות הקיים (`record_chair_feedback` → "עמדת ועדת הערר" ב-`analysis-and-research.md`
+  → `get_chair_directions` → `approve_direction`); לידים שנדחו נמחקים.
+- **הכותב צורך מקור-מעוגן בלבד ([INV-WR1–WR5](#4-invariants-של-התחום--תוכן-החלטה-מנומקת) +
+  [INV-LRN5](07-learning.md)):** הקלט לכתיבה הוא **פלט-המנתח המעוגן + ההנחיות-המאושרות** —
+  **לעולם לא** הלידים הגולמיים של שטן-מליץ. השכבה אינה מסלול-נתונים מקביל לכתיבה
+  ([G2](00-constitution.md#inv-g2-מקור-אמת-יחיד--אין-מסלולים-מקבילים-מתפצלים)) ואינה כותבת
+  שום שכבת-קול/ידע (INV-LRN5) — read-only ל-`critique-gemini.md` בלבד.
+- **מקור-אמת לזרימה:** [X4 INV-AG4](X4-agents.md#inv-ag4-שער-שטן-מליץ--red-team-לידים-לא-סמכותיים-תחת-אישור-יור)
+  + [legal-ceo.md](../../.claude/agents/legal-ceo.md) "שלב A2"/"שלב A3" +
+  [legal-analyst-gemini-critique.md](../../.claude/agents/legal-analyst-gemini-critique.md).
+  (שינוי-מדיניות יו"ר 2026-06-30: from on-demand to standing gate — TaskMaster `legal-ai` #211.)
+
 ---
 
 ## 2. ארכיטקטורת 12 הבלוקים (סיכום)

docs/spec/06-export.md

@@ -111,10 +111,14 @@ Model / Single Source of Truth) · SSOT (Single Source of Truth principle) | ס
 פלט מגורסן + idempotent מבחינת-תוכן; אוכף את
 [INV-G2](00-constitution.md#inv-g2-מקור-אמת-יחיד--אין-מסלולים-מקבילים-מתפצלים) וכלל-ההנדסה
 "סימטריה" (חוקה §6).
-**הפרה ידועה:** אחרי `revise_draft`/`apply_user_edit`, ה-DOCX המסומן `active_draft_path` הופך
-ל"מקור-האמת" לעריכות-Track-Changes העוקבות (`db.py:185-188`), ו**בלוקי-ה-DB אינם מתעדכנים
-חזרה** — הנתון-הנגזר זוחל למקור-אמת בפועל בלי סנכרון לאחור. **יעד:** או re-sync מהבלוקים, או
-חוזה מפורש ש-`active_draft_path` הוא רק אנקור-revision ולא מקור-תוכן → ראה [§5](#5-current-vs-target).
+**הפרה ידועה — ✅ מגודרת (FU-7, GAP-17):** אחרי `revise_draft`/`apply_user_edit` ה-DOCX
+המסומן `active_draft_path` משמש כאנקור לעריכות-Track-Changes העוקבות, ובלוקי-ה-DB אינם
+מתעדכנים חזרה — סטייה אפשרית בין הבלוקים למסמך-החי. **התיקון שנבחר (חוזה מפורש, לא re-sync):**
+`active_draft_path` הוא **אנקור-revision בלבד, לא מקור-תוכן מתחרה** — ה-DB נשאר מקור-האמת. סטייה
+נלכדת בדגל `cases.blocks_stale` (V22, `db.py:1142-1148`): `revise_draft`/`apply_user_edit`
+מסמנים `mark_blocks_stale(case_id, True)` (`drafting.py:688,789`) ו-`export_docx` מנקה
+(`mark_blocks_stale(case_id, False)`, `drafting.py:504`) — הדגל הוא ה-drift-detection שחושף מתי
+ה-DOCX-החי נסחף מהבלוקים. ראה [§5](#5-current-vs-target).
 
 ### INV-EX2: עקיבוּת-מקור נשמרת בהחלטה המיוצאת (→G9)
 **כלל:** ההחלטה המיוצאת **שומרת על עקיבוּת-מקור** היכן שנדרש — סמכויות-משפטיות מצוטטות
@@ -137,29 +141,53 @@ Lewis et al. (2020, NeurIPS — RAG attribution) | סטטוס: verified
 **מקורות:** NCSC/JTC — *Principles & Practices for AI Use in Courts* (controlled, auditable
 output) · Council of Europe / CEPEJ (2018, under user control) · Federal Judicial Center —
 *Judicial Writing Manual* (2d ed.) | סטטוס: verified
-**אכיפה:** `export_blocked = critical_failures > 0` (`qa_validator.py:362`); **נאכף ברמת-
-הזרימה/agent בלבד** — הסוכן `legal-exporter` מחויב להריץ `validate_decision` ולבדוק
-כשלים-קריטיים לפני ייצוא (`.claude/agents/legal-exporter.md:71,149`).
-**הפרה ידועה:** `export_docx` (`drafting.py:384`) **אינו** קורא ל-`validate_decision` בעצמו —
-הוא ניגש ישירות ל-`docx_exporter.export_decision` בלי לבדוק `export_blocked`. החסימה תלויה
-במשמעת-הסוכן ואינה hard-block בקוד-הייצוא → ראה [§5](#5-current-vs-target) (תואם
-[05-qa-review §4](05-qa-review.md#4-current-vs-target--ממצאי-audit)).
+**אכיפה — ✅ hard-block בקוד (FU-6, GAP-15):** `export_docx` (`drafting.py:462`) **בודק לעצמו**
+לפני כל ייצוא — `db.qa_run_exists` (אם QA לא רץ כלל → חסום) ו-`db.get_critical_qa_failures`
+(אם יש כשל-קריטי → חסום) — **לפני** הגישה ל-`docx_exporter.export_decision`. אלו `SELECT` זולים
+על ה-`qa_results` המאוחסנים (לא הרצת-LLM חוזרת). נוסף על `export_blocked = critical_failures > 0`
+ברמת-הזרימה ועל משמעת-הסוכן `legal-exporter` (`.claude/agents/legal-exporter.md:71,149`) — כך
+שאי-אפשר לעקוף את השער אפילו בקריאה ישירה ל-`export_docx`.
+**הפרה ידועה — ✅ נסגרה (FU-6, GAP-15):** בעבר `export_docx` ניגש ישירות ל-exporter בלי לבדוק
+`export_blocked` (אכוף-זרימה בלבד, ניתן-לעקיפה). נסגר ע"י ה-hard-block לעיל; מאומת ב-
+`test_export_qa_gate.py` (חסום ללא-QA · חסום בכשל-קריטי · עובר כשנקי).
+
+### INV-EX4: מקור-תבנית-יחיד — כל מפיק-מסמך מחיל סגנונות-טמפלט בלבד (→G2)
+**כלל:** **כל** מפיק של מסמך-Word הקשור-להחלטה — סופי (`export_docx`), טיוטת-ביניים
+(`export_interim_draft`), סיכום-מנהלים (`party-claims-summary`), ומסמך-הניתוח
+(`build_analysis_docx`), וכן מסלול-העריכה (`revise_draft`/`apply_user_edit` →
+`docx_reviser`) — **חייב לטעון את הטמפלט היחיד** `skills/docx/decision_template.docx`
+ולהחיל עיצוב **אך-ורק דרך שמות-הסגנונות שמוגדרים בו** (`paragraph.style = "…"` /
+`<w:pStyle w:val="…"/>`). **אסור hand-setting** של משפחת-גופן (`rFonts`), גודל-גופן
+(`sz`/`szCs`), או שוליים — אלה נגזרים מ-`styles.xml` של הטמפלט (cascade). היחיד שמותר
+לכפות מפורשות הוא **יישור-הגוף הקנוני `jc=both`** היכן שהוא דורס סגנון שמגיע ממורכז
+(`_set_paragraph_jc`, `docx_exporter.py:80`) — כי זו ההתנהגות-הקנונית, לא עיצוב-אד-הוק.
+סמני-כיווניות (`<w:bidi>`, run `<w:rtl/>`) מותרים (כיווניות, לא עיצוב). **טמפלט אחד = מקור-סגנון
+יחיד; אין שכבת-סגנון מקבילה.**
+**מקורות:** SSOT (Single Source of Truth) · Martin Fowler — Canonical Data Model · ECMA-376
+(OOXML) §17.7.2 — הורשת-סגנונות-פסקה (paragraph style inheritance / cascade) | סטטוס: verified
+**אכיפה:** `docx_exporter` ו-`analysis_docx_exporter` טוענים `TEMPLATE_PATH` ומחילים style-by-name
+(ראה §2); `docx_reviser` מחיל `pStyle` מן-הטמפלט + `jc=both` (לא פונט/גודל ידני). שומר-CI
+`test_docx_template_single_source.py` גורף את מודולי-מפיקי-ה-DOCX ונכשל על `rFonts`/`sz`/`szCs`
+ידני מחוץ לטוען-הטמפלט. אוכף את
+[INV-G2](00-constitution.md#inv-g2-מקור-אמת-יחיד--אין-מסלולים-מקבילים-מתפצלים) בשכבת-הרינדור.
+**הפרה ידועה — ✅ נסגרה:** `docx_reviser._build_paragraph` כפה `jc=right` + David/`sz` ידני על כל
+פסקה שהוכנסה במסלול-העריכה (`revise_draft`/`apply_user_edit`) → פלט מסלול-העריכה איבד את היישור
+הדו-צדדי של הטמפלט והכניס עיצוב-אד-הוק. נסגר ע"י החלת `pStyle` מן-הטמפלט; מאומת ב-
+`test_docx_reviser.py` + `test_docx_template_single_source.py`.
 
 ---
 
 ## 5. Current vs Target
 
-- **שער-ייצוא אכוף-זרימה ולא אכוף-קוד (INV-EX3 / INV-QA3).** אומת בקוד: `export_docx`
-  (`drafting.py:384-419`) קורא ישירות ל-`docx_exporter.export_decision` (`:403`) ללא קריאה
-  ל-`qa_validator.validate_decision` ובלי בדיקת `export_blocked`. החסימה מתקיימת רק כי הסוכן
-  `legal-exporter` מחויב להריץ QA קודם (`legal-exporter.md:71,149`) — אדם/סוכן שיקרא
-  ל-`export_docx` ישירות **יעקוף** את השער. **יעד:** hard-block בתוך `export_docx` — שליפת
-  `qa_results`/`export_blocked` ודחייה לפני כתיבת ה-DOCX, כך שאי-אפשר לעקוף.
-- **`active_draft_path` כ-derived-שזוחל-למקור (INV-EX1).** ה-DOCX נגזר מהבלוקים בייצוא הראשון,
-  אך אחרי עריכה (`revise_draft`/`apply_user_edit`) ה-DOCX הופך ל"מקור-האמת" לעריכות הבאות
-  (`db.py:185-188`) בלי לעדכן את `decision_blocks` חזרה — סטייה אפשרית בין הבלוקים למסמך-החי.
-  **יעד:** חוזה מפורש — או re-sync מהבלוקים, או הגדרת `active_draft_path` כאנקור-revision בלבד
-  (לא מקור-תוכן), עם בדיקת-בריאות לגילוי drift בין הבלוקים ל-DOCX הפעיל.
+- **שער-ייצוא — ✅ hard-block בקוד (INV-EX3 / INV-QA3, FU-6 / GAP-15).** `export_docx`
+  (`drafting.py:462-494`) בודק `db.qa_run_exists` + `db.get_critical_qa_failures` ודוחה לפני
+  הגישה ל-`docx_exporter.export_decision` — לא ניתן לעקוף בקריאה ישירה. מאומת ב-
+  `test_export_qa_gate.py` (3 מקרים: ללא-QA / כשל-קריטי / נקי).
+- **`active_draft_path` כ-derived (INV-EX1, FU-7 / GAP-17) — ✅ מגודר בחוזה מפורש.** ה-DB נשאר
+  מקור-האמת; `active_draft_path` הוא אנקור-revision בלבד. ה-drift בין הבלוקים ל-DOCX-החי נלכד
+  בדגל `cases.blocks_stale`: נדלק ב-`revise_draft`/`apply_user_edit` (`drafting.py:688,789`),
+  כובה ב-`export_docx` (`drafting.py:504`). **שארית (low-pri):** health-check שמתריע על
+  `blocks_stale=true` עתיק — תיעוד-המשך, לא חוסם.
 
 ---
 

docs/spec/07-learning.md

@@ -73,6 +73,8 @@
 
 **הקריאה:** טבלת `style_distance_history` (append-only) + `GET /api/learning/style-distance-history`. ירידה ב-`anti_pattern_total`/`change_percent` ככל ש-N גדל = **הוכחה מתגלגלת שהלמידה מכלילה** (INV-LRN4 — זהו משטח-המגמה של "ניגוד-האמת"). **אזהרת-פרשנות:** `change_percent` מערבב סגנון עם שלמות-תוכן (לפעמים היו"ר מכפילה אורך כי חסרה מהות, לפעמים חותכת) → `anti_pattern_total` הוא הסיגנל הנקי-יותר לסגנון. שימוש-חוזר בשירות `style_distance` ובבריכת `appeal_type_rules` — אין מסלול-מדד מקביל (G2).
 
+> **המלצת-הכיול style-ranked (#213):** `recommend_effort` ב-`scripts/calibrate_effort.py` מדרג efforts לפי `anti_pattern_total` **ראשי** (סיגנל-סגנון נקי) → סטיית-יחסי-זהב → `distance` (tiebreak קלוש). `change_percent` **מדווח-לא-מדורג** מאותה אזהרת-פרשנות לעיל (מערבב סגנון עם שלמות-תוכן). ההמלצה נושאת דגל-`confidence` ("clear"/"weak") כך שבחירה שההובלה-ב-anti שלה בתוך-הרעש (epsilon=max(0.5, 0.20·spread)) לא מוצגת כוודאית. אין מסלול-מדד מקביל — שימוש-חוזר ב-`style_distance` (G2).
+
 > **חלון נקי חד-פעמי:** תיק שכבר `lessons_folded` פספס את חלון ה-held-out שלו — אין backfill. הטבלה מתמלאת קדימה מהסופי הבא.
 
 ---

docs/spec/X4-agents.md

@@ -59,6 +59,16 @@ Paperclip בקונפליקט (project-specific מנצח default), אך אינו
   ל-`SKILL.md`/`lessons.md` — מופע של [G10](00-constitution.md#inv-g10-המערכת-מסייעת--שערים-אנושיים-הם-invariant).
 - **company_id פר-סוכן.** כל שורה בטבלה מיוצגת פעמיים (CMP + CMPA); ה-CEO לכל חברה שונה
   ([X2 §1](X2-multi-company.md)). הסוכן פועל רק בטווח-החברה שלו ([X2 §2](X2-multi-company.md)).
+- **שטן-מליץ (Gemini red-team) — שער-קבע אחרי הניתוח, לא חלק מ-7-הדומייניים.** סוכן
+  `legal-analyst-gemini-critique` (`gemini_local`, CMP+CMPA) רץ **אחרי** שלב-הניתוח של ה-Case
+  Analyst (Opus) ו**לפני** הכותב, כ**שער-קבע (standing gate)** שה-CEO מפעיל **אוטומטית** — לא
+  on-demand. הוא read-only ומפיק `critique-gemini.md` = **מזכר-לידים לא-סמכותי**. ה-CEO **עוצר
+  את הזרימה לאישור-יו"ר של הלידים** (ה-issue הראשי ל-`in_review`); רק לידים שהיו"ר מאשר מומרים
+  ל-`chair_directions` (דרך מנגנון-ההנחיות הקיים — `record_chair_feedback` → `get_chair_directions`
+  → `approve_direction`), והכותב צורך **רק** את פלט-המנתח המעוגן + ההנחיות-המאושרות — **לעולם
+  לא** את הלידים הגולמיים. מקיים [INV-AG4](#inv-ag4-שער-שטן-מליץ--red-team-לידים-לא-סמכותיים-תחת-אישור-יור) +
+  [04-analysis-writing §1.6](04-analysis-writing.md). מקור-אמת לזרימה: [legal-ceo.md](../../.claude/agents/legal-ceo.md)
+  "שלב A2"/"שלב A3" + [legal-analyst-gemini-critique.md](../../.claude/agents/legal-analyst-gemini-critique.md).
 
 ### 2א. מפת-הרשאות (tool grants) — frontmatter מול הוראות
 
@@ -140,6 +150,25 @@ another company`, [X2 §2](X2-multi-company.md)).
 **אכיפה:** בדיקת-עקביות tools↔instructions (FU-13 ✅ 2026-06-06). אכיפה אוטומטית עתידית — בתת-פרויקט 5 (spec-guardian).
 **הפרה ידועה:** —  (טופל ב-FU-13: legal-analyst קיבל `aggregate_claims_to_arguments`; researcher כבר היה תקין; `extract_references`/`extract_internal_citations` הם מטלת-researcher, לא analyst — ראה §2א).
 
+### INV-AG4: שער שטן-מליץ — red-team לידים לא-סמכותיים תחת אישור-יו"ר
+**כלל:** אחרי שלב-הניתוח (`analysis-and-research.md` תקין) וב**לפני** הפעלת הכותב, ה-CEO מפעיל
+**אוטומטית** את סוכן שטן-מליץ (Gemini red-team, read-only) כ**שער-קבע** — לא on-demand. הפלט
+(`critique-gemini.md`) הוא **מזכר-לידים לא-סמכותי**; ה-CEO **עוצר את הזרימה לעצירת-אישור קשיחה
+של היו"ר** (issue ראשי ל-`in_review` + מייל) ואינו מתקדם לכותב בלי הכרעת-יו"ר מפורשת. **רק לידים
+שהיו"ר אישר** מומרים ל-`chair_directions` דרך מנגנון-ההנחיות הקיים (`record_chair_feedback`
+→ `get_chair_directions` → `approve_direction`); לידים שנדחו נמחקים. הכותב צורך **אך-ורק** את
+פלט-המנתח המעוגן + ההנחיות-המאושרות — **לעולם לא** את הלידים הגולמיים. מופע של
+[G10](00-constitution.md#inv-g10-המערכת-מסייעת--שערים-אנושיים-הם-invariant) (שער אנושי
+לא-עקיף) ושל [INV-AH](../anti-hallucination-gate.md) / [INV-LRN5](07-learning.md) (לידים
+לא-סמכותיים, אינם מוזנים אוטומטית לקול/למהות).
+**מקור-סמכות:** [legal-ceo.md](../../.claude/agents/legal-ceo.md) ("שלב A2"/"שלב A3" + "מפת
+סטטוסים") + [legal-analyst-gemini-critique.md](../../.claude/agents/legal-analyst-gemini-critique.md)
++ [HEARTBEAT.md §7](../../.claude/agents/HEARTBEAT.md). (invariant פרויקטלי-תפעולי — ללא
+פרוטוקול ≥3-המקורות; משרת את G10 + INV-AH/INV-LRN5.)
+**אכיפה:** פרוצדורלית (נוהל ה-CEO — "אל תמשיך לכותב בלי `critique-gemini.md` + אישור-יו"ר ללידים")
++ עצירת-`in_review` של ה-issue הראשי; אין שער-קוד אוטומטי (כמו יתר ה-INV-AG*).
+**הפרה ידועה:** — (שינוי-מדיניות יו"ר 2026-06-30: שטן-מליץ עבר מ-on-demand ל-שער-קבע; TaskMaster `legal-ai` #211).
+
 ---
 
 ## 5. חיווט הספ לסוכנים — בוצע (FU-8b)

docs/spec/X6-ui-api-contract.md

@@ -105,13 +105,36 @@ TanStack Query — *Important Defaults* (staleTime/refetch) (https://tanstack.co
 **אכיפה:** `halacha_backlog` מרונדר ב-/operations (לא נזרק); `findings_approved` (review_status, צרכן אמיתי)
 החליף את `findings_applied` (דגל מת); מוני-סוכנים מסמנים "חלקי" כשחברה לא-נטענה (גל-1 #130).
 
+### INV-UI9: זמן-מאוחסן ב-UTC · תצוגה אנושית דטרמיניסטית ב-Asia/Jerusalem
+**כלל:** **האחסון נשאר UTC** מקצה-לקצה — Postgres `TIMESTAMPTZ`, וה-API מחזיר ISO-UTC; אין נרמול-tz
+בשכבת-האחסון/ה-API. **שכבת-התצוגה בלבד** ממירה, ובאופן **דטרמיניסטי**: כל חותמת-זמן אנושית מרונדרת
+ב-`timeZone: "Asia/Jerusalem"` (locale `he-IL`), בלי תלות ב-tz של ה-host/קונטיינר/דפדפן. כיוון
+שהקונטיינר רץ ב-UTC, פירמוט ad-hoc (`new Date(iso).toLocaleDateString("he-IL")`, `.getHours()`,
+`toISOString().slice(0,10)`) מציג UTC תחת SSR (ושוב UTC בכל runtime לא-ישראלי) → drift. נעיצת
+`Asia/Jerusalem` מבטיחה ש-SSR והדפדפן מסכימים. **פורמט-התצוגה (צורת התאריך/שעה) נשמר זהה — רק ה-tz
+מנורמל.** מופע של [G2](00-constitution.md#inv-g2-מקור-אמת-יחיד--אין-מסלולים-מקבילים-מתפצלים)
+(מפרמט-תאריך **יחיד**, אין פירמוט-זמן מקביל-מפוזר) בשכבת-התצוגה.
+**מקורות:** ECMAScript Internationalization API — `Intl.DateTimeFormat` `timeZone` option
+(https://tc39.es/ecma402/#datetimeformat-objects) · MDN — *Intl.DateTimeFormat* (IANA time zones)
+(https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Intl/DateTimeFormat) ·
+Next.js — *Rendering* (SSR משתמש בשעון-השרת אם לא ננעץ)
+(https://nextjs.org/docs/app/building-your-application/rendering) | סטטוס: verified
+**אכיפה:** מפרמט-תאריך יחיד ב-[web-ui/src/lib/format-date.ts](../../web-ui/src/lib/format-date.ts)
+(`formatDate`/`formatDateShort`/`formatDateLong`/`formatDateTime`/`formatDateTimeFull`/`formatTime`/
+`formatIsoDate`/`formatRelative` + עוזרי-Israel `getIsraelYear`/`israelDayKey`/`israelMidnightMs`/
+`israelParts`), כולם נעוצים ל-`Asia/Jerusalem`; כל משטחי-ה-UI מייבאים ממנו ולא מפרמטים tz ad-hoc.
+פירמוט-**מספרים** (`n.toLocaleString`) מחוץ-לתחום (אינו חותמת-זמן). **חריג מוצהר:** קלט
+`<input type="datetime-local">` (בקרת-burst ב-`/operations`) הוא ערך הניתן-לכתיבה ע"י המפעיל, לא תצוגת-
+מאוחסן — נשאר בנייה מקומית. **אודיט-נלווה (לא מיושם):** הגדרת `TZ=Asia/Jerusalem` בקונטיינר בטוחה רק
+אם אף לוגיקה עסקית אינה תלויה בשעון-ה-process כ-UTC.
+
 ---
 
 ## 3. כללי-עיצוב (Design Rules) — נגזרים מה-invariants
 - **SSoT ל-enums/תוויות/tones:** כל enum (CaseStatus, PracticeArea, AppealSubtype, DocType, outcome) +
   תוויותיו + צבעיו מוגדרים **פעם אחת** ונצרכים מיבוא — לא משוכפלים בין דפים/רכיבים (מופע UI1/G2).
-- **helpers משותפים:** פירמוט-תאריך, builder ל-FormData (העלאות), רכיב-שגיאה, query-config (intervals) —
-  משותפים, לא מועתקים.
+- **helpers משותפים:** פירמוט-תאריך (מפרמט-יחיד נעוץ-Asia/Jerusalem ב-[format-date.ts](../../web-ui/src/lib/format-date.ts),
+  INV-UI9), builder ל-FormData (העלאות), רכיב-שגיאה, query-config (intervals) — משותפים, לא מועתקים.
 - **חוזי-טופס:** ראה INV-UI6 ([X8](X8-field-provenance.md)).
 
 הממצאים הקונקרטיים (כפילויות, הגדרות-שגויות, redundancy) ב-[ui-audit.md](ui-audit.md); התיקון — **FU-10**.

docs/spec/gap-audit.md

@@ -232,3 +232,25 @@ FU-7 אחרי FU-1; FU-5 אחרי FU-2; FU-8 אחרי ייצוב-הספ. **(מח
 אחרי FU-1. FU-11 (מילוי-שדות) עצמאי. FU-13 (סוכנים+skills) אחרי ייצוב-הספ.
 **סיווג:** pure-code — FU-9/10/11/13/14; +data-migration קל — FU-12; +chair-decision — FU-15 (rotation סודות).
 priority בפועל — של היו"ר.
+
+---
+
+## ✅ ווידוא-סגירה — 7 דליפות עיצוב-מחדש זרימת-העבודה (WS7, #207, 2026-06-30)
+
+תת-פרויקט עיצוב-מחדש זרימת-העבודה (תוכנית `groovy-doodling-token`, WS7) דרש סגירת **7 דליפות
+ידועות** תוך-כדי. ביקורת חוצת-קוד (#207) אימתה שכל ה-7 כבר **נסגרו במחזורים קודמים** (בעיקר
+FU-1/4/5/6/7 + גלי-ה-IA), ושהספ נשא תיאורי-הפרה מיושנים. ה-PR מיישר את הספ למצב-הקוד בפועל
+ומאמת כל סגירה בבדיקה. **אין כאן GAP/FU חדש — זו ביקורת-יישור, לא פתרון-מחדש (G2).**
+
+| # | דליפה | invariant | סטטוס | FU | סגירה (קוד) | בדיקה |
+|---|-------|-----------|-------|-----|-------------|-------|
+| 1 | הלכות חוצות-קורפוס (#56) | INV-RET1/G5 (GAP-10) | ✅ סגור | FU-4 | `cl.source_kind` ב-`halacha_filters` בשתי הפונקציות (`db.py:7516,7519`) | `test_precedent_corpus_isolation.py` |
+| 2 | backlog-הלכות בלתי-נראה | INV-QA1/G10 (GAP-14) | ✅ סגור | FU-5 | health `halacha_backlog` (`app.py:2352-2364`) + כלי `extraction_status`/`halachot_pending` | — (health-endpoint) |
+| 3 | אין eval-harness | INV-RET4/G8 (GAP-11) | ✅ סגור | FU-5 | `scripts/eval_retrieval.py` (P/R/MRR/nDCG מול gold-set+baseline) | `eval_retrieval.py --self-test` ("ALL PASS") |
+| 4 | שער-ייצוא לא hard-block | INV-EX3/QA3/G10 (GAP-15) | ✅ סגור | FU-6 | `export_docx` בודק `qa_run_exists`+`get_critical_qa_failures` לפני exporter (`drafting.py:462-494`) | `test_export_qa_gate.py` |
+| 5 | מטא במסלול-פנימי | INV-ING3/G2 (GAP-02) | ✅ סגור | FU-1 | מסלול קנוני `ingest.ingest_document` מתזמן metadata+halacha יחד (`ingest.py:233-234`) | `test_unified_ingest.py` |
+| 6 | DOCX זוחל למקור-אמת | INV-EX1/G2 (GAP-17) | ✅ מגודר | FU-7 | חוזה: `active_draft_path`=אנקור בלבד; drift בדגל `cases.blocks_stale` (V22) | (חוזה+דגל; שארית health-check low-pri) |
+| 7 | cache-invalidation ב-UI | INV-UI5/IA (GAP-33) | ✅ סגור | FU-10/IA | `qc.invalidateQueries` על מוטציות (`web-ui/src/lib/api/*.ts`) + גלי-IA #130-132 | (TanStack Query) |
+
+**שאריות (low-pri, מתועדות לא-חוסמות):** דליפה #6 — health-check שמתריע על `blocks_stale=true`
+עתיק (הדגל קיים ונאכף; חסר רק התרעת-נראות). אינה חוסמת את WS7; ניתן ל-FU-7-המשך.

mcp-server/src/legal_mcp/case_status_model.py

@@ -0,0 +1,118 @@
+"""Single source of truth for the case-status lifecycle (the *status model*).
+
+Every consumer derives its order / labels / phase from this one registry:
+  • ``CaseStatus`` enum (models.py) — the type-level key set (kept consistent
+    by the assertion at the bottom of this module);
+  • the forward-only ``STATUS_ORDER`` guard in ``tools/cases.py``;
+  • the ``GET /api/status-model`` endpoint (web/app.py) that the frontend
+    mirror (``web-ui/src/lib/api/case-status.ts``) is generated against.
+
+To make a status *do something* on entry (notify, kick a job, …), set its
+``on_enter`` to an action key — the single place where a status's behaviour is
+declared. ``tools/cases.py`` dispatches it when a case transitions into that
+status (forward-only), so the behaviour lives next to the definition.
+
+The five **phases** are the coarse pipeline the 12 statuses collapse into for the
+header stepper. Intermediate analyst/research states (``analyst_verified``,
+``research_complete``) are first-class canonical statuses here — the agents set
+them, so they must be in the model rather than fall between the canonical set
+and the legacy bucket (the bug that made the chip, stepper and manual-changer
+disagree).
+"""
+
+from __future__ import annotations
+
+from dataclasses import dataclass
+
+# Ordered 5-phase pipeline (key → Hebrew label) — coarse view of the lifecycle.
+PHASES: list[tuple[str, str]] = [
+    ("intake", "קליטה ועיבוד"),
+    ("prep", "הכנת תיק"),
+    ("thinking", "ניתוח וכיוון"),
+    ("writing", "כתיבת טיוטה"),
+    ("done", "סגירה"),
+]
+PHASE_LABELS: dict[str, str] = dict(PHASES)
+
+
+@dataclass(frozen=True)
+class StatusDef:
+    key: str
+    label: str            # Hebrew label (chip / dropdown)
+    description: str       # Hebrew one-liner (status guide)
+    phase: str            # one of PHASES keys
+    selectable: bool = True   # offered in the manual status-changer dropdown
+    terminal: bool = False
+    on_enter: str | None = None  # future: action key dispatched on entry
+
+
+# THE lifecycle — ordered. STATUS_ORDER, the enum and the frontend all derive
+# from this list. Insert intermediates in workflow order.
+STATUS_DEFS: list[StatusDef] = [
+    StatusDef("new", "חדש", "התיק נוצר וממתין להעלאת מסמכים", "intake"),
+    StatusDef("processing", "בעיבוד", "המערכת מעבדת ומנתחת את המסמכים", "intake"),
+    StatusDef("documents_ready", "מסמכים מוכנים", "כל המסמכים עובדו ומוכנים לעבודה", "prep"),
+    StatusDef("analyst_verified", "ניתוח אומת", "המנתח סיים ואימת את הניתוח — ממתין להכרעת תוצאה של היו״ר", "thinking"),
+    StatusDef("research_complete", "מחקר הושלם", "חקר התקדימים הושלם (מסלול נפרד מהמנתח)", "thinking"),
+    StatusDef("outcome_set", "תוצאה נקבעה", "נקבעה תוצאה צפויה לערר", "thinking"),
+    StatusDef("direction_approved", "כיוון אושר", "כיוון ההחלטה אושר — בהעמקת ניתוח וכתיבה", "thinking"),
+    StatusDef("qa_review", "בדיקת איכות", "הטיוטה בבדיקת איכות אוטומטית", "writing"),
+    StatusDef("drafted", "טיוטה", "טיוטה מוכנה לעיון", "writing"),
+    StatusDef("exported", "יוצא", "ההחלטה יוצאה לקובץ DOCX", "done"),
+    StatusDef("reviewed", "נבדק", 'ההחלטה נבדקה ע"י היו"ר', "done"),
+    StatusDef("final", "סופי", "החלטה סופית — מוכנה להגשה", "done", terminal=True),
+]
+
+BY_KEY: dict[str, StatusDef] = {d.key: d for d in STATUS_DEFS}
+STATUS_ORDER: list[str] = [d.key for d in STATUS_DEFS]
+
+
+def phase_of(status: str | None) -> str | None:
+    """The pipeline phase a status belongs to (None for unknown values)."""
+    d = BY_KEY.get(status or "")
+    return d.phase if d else None
+
+
+def label_of(status: str | None) -> str:
+    d = BY_KEY.get(status or "")
+    return d.label if d else (status or "")
+
+
+def to_dict() -> dict:
+    """Serialisable status model for ``GET /api/status-model`` (frontend SSoT)."""
+    return {
+        "statuses": [
+            {
+                "key": d.key,
+                "label": d.label,
+                "description": d.description,
+                "phase": d.phase,
+                "selectable": d.selectable,
+                "terminal": d.terminal,
+                "on_enter": d.on_enter,
+            }
+            for d in STATUS_DEFS
+        ],
+        "phases": [{"key": k, "label": v} for k, v in PHASES],
+    }
+
+
+# Drift guard — the registry IS the canonical key set; the CaseStatus enum must
+# match it exactly. Importing here (models.py never imports this module) is safe.
+def _assert_consistent() -> None:
+    from legal_mcp.models import CaseStatus
+
+    registry = {d.key for d in STATUS_DEFS}
+    enum_keys = {s.value for s in CaseStatus}
+    if registry != enum_keys:
+        raise RuntimeError(
+            "case_status_model drift: registry vs CaseStatus enum differ — "
+            f"only-in-registry={registry - enum_keys}, only-in-enum={enum_keys - registry}"
+        )
+    phase_keys = {k for k, _ in PHASES}
+    bad = {d.key: d.phase for d in STATUS_DEFS if d.phase not in phase_keys}
+    if bad:
+        raise RuntimeError(f"case_status_model: statuses with unknown phase: {bad}")
+
+
+_assert_consistent()

mcp-server/src/legal_mcp/models.py

@@ -9,13 +9,18 @@ from uuid import UUID
 from pydantic import BaseModel, Field
 
 
-# Core case lifecycle — kept in sync with STATUS_ORDER in tools/cases.py and the
-# frontend SSoT web-ui/src/lib/api/case-status.ts. Trimmed from 17 → 10 (the
-# decorative mid-stage markers that no pipeline code ever set were removed).
+# Core case lifecycle. The canonical key set, order, labels, phases and (future)
+# per-status actions all live in ONE place — legal_mcp/case_status_model.py —
+# which asserts this enum matches it. The frontend mirror is
+# web-ui/src/lib/api/case-status.ts (generated against GET /api/status-model).
+# The analyst/research intermediate states are first-class canonical statuses
+# (the agents set them) — not legacy.
 class CaseStatus(str, enum.Enum):
     NEW = "new"
     PROCESSING = "processing"
     DOCUMENTS_READY = "documents_ready"
+    ANALYST_VERIFIED = "analyst_verified"
+    RESEARCH_COMPLETE = "research_complete"
     OUTCOME_SET = "outcome_set"
     DIRECTION_APPROVED = "direction_approved"
     QA_REVIEW = "qa_review"

mcp-server/src/legal_mcp/server.py

@@ -759,13 +759,15 @@ async def get_appraiser_facts(case_number: str) -> str:
 # ── Protocol comparative analysis (WS4 / #203) — ניתוח פרוטוקול ────
 
 @mcp.tool()
-async def analyze_protocol(case_number: str) -> str:
+async def analyze_protocol(case_number: str, document_id: str = "") -> str:
     """ניתוח השוואתי של פרוטוקול-דיון מול כתבי-הטענות: אילו טענות ירדו/חוזקו/עלו-חדשות + חידוד-שאלות + חילוץ א–ד.
 
-    מזין ל"ידע-התיק" (protocol_analysis); דורש פרוטוקול doc_type='protocol' +
-    טיעונים מאוגדים. Claude מקומי Opus 4.8 effort=high; re-run מחליף (idempotent).
+    בוחר אוטומטית פרוטוקול ועדת-ערר (scope שאינו 'lower'); document_id מכוון לדיון
+    מדויק כשיש כמה פרוטוקולים. מזין ל"ידע-התיק" (protocol_analysis); דורש פרוטוקול
+    doc_type='protocol' + טיעונים מאוגדים. Claude מקומי Opus 4.8 effort=high;
+    re-run מחליף (idempotent).
     """
-    return await drafting.analyze_protocol(case_number)
+    return await drafting.analyze_protocol(case_number, document_id=document_id)
 
 
 @mcp.tool()

mcp-server/src/legal_mcp/services/argument_aggregator.py

@@ -31,6 +31,10 @@ logger = logging.getLogger(__name__)
 
 # Allowed enum values mirror the DB CHECK constraints.
 ALLOWED_PARTIES = {"appellant", "respondent", "committee", "permit_applicant", "unknown"}
+# Sides that may comprise multiple distinct litigants with potentially opposing
+# positions — aggregated per source pleading (party_name) rather than collapsed
+# into one bucket (#224). Appellant/committee speak with a single voice.
+SPLIT_PARTIES = {"respondent", "permit_applicant"}
 ALLOWED_PRIORITIES = {"threshold", "substantive", "procedural", "relief"}
 
 # Hebrew labels for the prompt (Claude needs context in the same
@@ -81,7 +85,7 @@ AGGREGATE_PROMPT_TEMPLATE = """אתה מנתח כתבי טענות בתחום ת
 """
 
 
-def _build_prompt(party: str, propositions: list[dict]) -> str:
+def _build_prompt(party: str, propositions: list[dict], party_name: str = "") -> str:
     """Compose the per-party aggregation prompt."""
     n = len(propositions)
     # Conservative target: ~1 argument per 2-3 propositions, clamped 4-12.
@@ -89,6 +93,11 @@ def _build_prompt(party: str, propositions: list[dict]) -> str:
     target_max = max(target_min + 1, min(12, n // 2 + 1))
 
     party_he = PARTY_LABELS_HE.get(party, party)
+    # For a split side (e.g. a specific respondent brief), name the brief so
+    # Claude aggregates only that litigant's position and does not conflate it
+    # with a co-respondent's separate — possibly opposing — pleading (#224).
+    if party_name:
+        party_he = f"{party_he} — {party_name}"
     # Strip noise from propositions for the prompt — Claude only needs
     # the id and the text to do the grouping.
     compact = [
@@ -139,12 +148,17 @@ def _normalize_argument(raw: dict, fallback_topic: str = "") -> dict | None:
 
 
 async def _aggregate_party(
-    party: str, propositions: list[dict],
+    party: str, propositions: list[dict], party_name: str = "",
 ) -> list[dict]:
-    """Ask Claude to group one party's propositions; return normalized rows."""
+    """Ask Claude to group one party's propositions; return normalized rows.
+
+    ``party_name`` names the specific pleading when this is a split side
+    (respondent / permit_applicant brief), so the prompt scopes to that
+    litigant's position only (#224).
+    """
     if not propositions:
         return []
-    prompt = _build_prompt(party, propositions)
+    prompt = _build_prompt(party, propositions, party_name=party_name)
 
     try:
         raw_result = await claude_session.query_json(prompt, tools="")  # no tool_use → no error_max_turns
@@ -207,7 +221,7 @@ async def aggregate_claims_to_arguments(
 
         # Pull all claims for this case, grouped by party.
         rows = await conn.fetch(
-            """SELECT id, party_role, claim_text, claim_index, source_document
+            """SELECT id, party_role, claim_text, claim_index, source_document, party_name
                FROM claims
                WHERE case_id = $1
                ORDER BY party_role, claim_index""",
@@ -221,14 +235,28 @@ async def aggregate_claims_to_arguments(
             "total": 0,
         }
 
-    # Group propositions by party.
-    by_party: dict[str, list[dict]] = {}
+    # Group propositions by party — and, for the multi-party sides (respondent /
+    # permit_applicant), further by their source pleading so opposing joint
+    # briefs (e.g. "כתב תשובה משיבות 2-3" vs "משיבים 4-6") aggregate into
+    # SEPARATE argument sets instead of collapsing into one "respondent" bucket
+    # (#224). Appellant/committee stay single-group (party_name=""). Splitting
+    # the large respondent set per-brief also keeps each Claude call small enough
+    # to succeed — a single 100+ proposition call previously returned non-JSON
+    # and silently dropped the whole side.
+    by_group: dict[tuple[str, str], list[dict]] = {}
     for r in rows:
         party = r["party_role"]
         # Map deprecated 'appraiser' or unknown labels to 'unknown'.
         if party not in ALLOWED_PARTIES:
             party = "unknown"
-        by_party.setdefault(party, []).append(dict(r))
+        # Prefer the stored party_name (stamped by the extractor, #224); fall
+        # back to source_document for legacy claims predating the stamping.
+        # Single-voice sides (appellant/committee) stay ''.
+        if party in SPLIT_PARTIES:
+            party_name = (r["party_name"] or "").strip() or (r["source_document"] or "").strip()
+        else:
+            party_name = ""
+        by_group.setdefault((party, party_name), []).append(dict(r))
 
     # Valid claim_ids for this case == the ids of the claims we just fetched.
     # The LLM is asked to echo back supporting claim_ids, but it may hallucinate
@@ -243,9 +271,11 @@ async def aggregate_claims_to_arguments(
     inserted = 0
     errors: list[str] = []
 
-    for party, props in by_party.items():
+    for (party, party_name), props in by_group.items():
+        # Display key for the per-side summary: keep opposing briefs distinct.
+        group_key = f"{party}·{party_name}" if party_name else party
         try:
-            arguments = await _aggregate_party(party, props)
+            arguments = await _aggregate_party(party, props, party_name=party_name)
         except RuntimeError as e:
             # Most likely cause: Claude CLI not installed (running from
             # the container). Don't crash — record the gap and continue.
@@ -259,11 +289,11 @@ async def aggregate_claims_to_arguments(
                     ),
                     "total": 0,
                 }
-            errors.append(f"{party}: {msg}")
+            errors.append(f"{group_key}: {msg}")
             continue
 
         if not arguments:
-            party_counts[party] = 0
+            party_counts[group_key] = 0
             continue
 
         async with pool.acquire() as conn:
@@ -271,12 +301,13 @@ async def aggregate_claims_to_arguments(
                 for idx, arg in enumerate(arguments):
                     arg_id = await conn.fetchval(
                         """INSERT INTO legal_arguments
-                            (case_id, party, argument_index, argument_title,
-                             argument_body, legal_topic, priority)
-                           VALUES ($1, $2, $3, $4, $5, $6, $7)
+                            (case_id, party, party_name, argument_index,
+                             argument_title, argument_body, legal_topic, priority)
+                           VALUES ($1, $2, $3, $4, $5, $6, $7, $8)
                            RETURNING id""",
                         case_id,
                         party,
+                        party_name,
                         idx + 1,
                         arg["title"],
                         arg["body"],
@@ -313,7 +344,7 @@ async def aggregate_claims_to_arguments(
                                 cid, arg_id, e,
                             )
                     inserted += 1
-        party_counts[party] = len(arguments)
+        party_counts[group_key] = len(arguments)
 
     result: dict = {
         "status": "completed",
@@ -338,9 +369,9 @@ async def get_legal_arguments(
     async with pool.acquire() as conn:
         if party and party in ALLOWED_PARTIES:
             rows = await conn.fetch(
-                """SELECT id, case_id, party, argument_index, argument_title,
-                          argument_body, legal_topic, priority, cited_precedents,
-                          created_at, updated_at
+                """SELECT id, case_id, party, party_name, argument_index,
+                          argument_title, argument_body, legal_topic, priority,
+                          cited_precedents, created_at, updated_at
                    FROM legal_arguments
                    WHERE case_id = $1 AND party = $2
                    ORDER BY priority, argument_index""",
@@ -348,12 +379,12 @@ async def get_legal_arguments(
             )
         else:
             rows = await conn.fetch(
-                """SELECT id, case_id, party, argument_index, argument_title,
-                          argument_body, legal_topic, priority, cited_precedents,
-                          created_at, updated_at
+                """SELECT id, case_id, party, party_name, argument_index,
+                          argument_title, argument_body, legal_topic, priority,
+                          cited_precedents, created_at, updated_at
                    FROM legal_arguments
                    WHERE case_id = $1
-                   ORDER BY party, priority, argument_index""",
+                   ORDER BY party, party_name, priority, argument_index""",
                 case_id,
             )
 

mcp-server/src/legal_mcp/services/audit.py

@@ -7,7 +7,7 @@ from __future__ import annotations
 
 import json
 import logging
-from datetime import datetime
+from datetime import datetime, timezone
 from uuid import UUID, uuid4
 
 from legal_mcp.services import db
@@ -38,7 +38,7 @@ async def log_action(
                VALUES ($1, $2, $3, $4, $5, $6, $7)""",
             uuid4(), action, case_id, document_id,
             json.dumps(details or {}, ensure_ascii=False, default=str),
-            user, datetime.utcnow(),
+            user, datetime.now(timezone.utc),
         )
     logger.info("AUDIT: %s | case=%s | user=%s | %s", action, case_id, user,
                 json.dumps(details or {}, ensure_ascii=False)[:200])

mcp-server/src/legal_mcp/services/block_writer.py

@@ -58,6 +58,18 @@ logger = logging.getLogger(__name__)
 # Output token note (Anthropic): Opus 4.8 supports large outputs; streaming is
 # handled by the CLI. `max_tokens` is advisory context for callers, not sent.
 GENERATION_MODEL = "claude-opus-4-8"  # single pinned model for every AI block (#204)
+# 1M-context build of the SAME pinned model (#216). The default build runs a
+# 200K-token context (~400K Hebrew chars at ~2 chars/token); large prompts —
+# notably block-yod, which carries the full case as source-context — overflow it.
+# Opus 4.8 offers a 1M-token window at standard pricing; `claude -p` exposes it
+# via the `[1m]` model id directly (no beta header — verified CLI 2.1.196). We
+# escalate to it only when the prompt is large, so small blocks stay on the
+# cheaper/faster 200K build. NOT a parallel model (G2) — same model, wider window.
+GENERATION_MODEL_1M = "claude-opus-4-8[1m]"
+# Escalate to the 1M build above this prompt size. Set below the 200K-token wall
+# (~400K chars) with headroom for output tokens (max_tokens up to 16K) + tokenizer
+# variance, so we switch before the standard build can overflow.
+_CTX_1M_THRESHOLD_CHARS = 350_000
 
 BLOCK_CONFIG = {
     "block-alef": {"index": 1, "title": "כותרת מוסדית", "gen_type": "template-fill", "model": "script"},
@@ -65,7 +77,7 @@ BLOCK_CONFIG = {
     "block-gimel":{"index": 3, "title": "צדדים", "gen_type": "template-fill", "model": "script"},
     "block-dalet":{"index": 4, "title": "החלטה", "gen_type": "template-fill", "model": "script"},
     "block-he":   {"index": 5, "title": "פתיחה", "gen_type": "paraphrase", "model": "ai", "effort": "medium", "max_tokens": 4096},
-    "block-vav":  {"index": 6, "title": "רקע עובדתי", "gen_type": "reproduction", "model": "ai", "effort": "medium", "max_tokens": 16384},
+    "block-vav":  {"index": 6, "title": "רקע עובדתי", "gen_type": "reproduction", "model": "ai", "effort": "low", "max_tokens": 16384},
     "block-zayin":{"index": 7, "title": "טענות הצדדים", "gen_type": "paraphrase", "model": "ai", "effort": "high", "max_tokens": 16384},
     "block-chet": {"index": 8, "title": "הליכים", "gen_type": "reproduction", "model": "ai", "effort": "medium", "max_tokens": 8192},
     "block-tet":  {"index": 9, "title": "תכניות חלות", "gen_type": "guided-synthesis", "model": "ai", "effort": "high", "max_tokens": 16384},
@@ -180,16 +192,16 @@ BLOCK_PROMPTS = {
 ## כללים קריטיים:
 - **סנתז טענות דומות** — אל תרשום כל טענה בנפרד. קבץ טענות דומות לנושא אחד. למשל: כל הטענות על הודעות → סעיף אחד, כל הטענות על רכוש משותף → סעיף אחד.
 - גוף שלישי: "העוררים טוענים כי...", "הוועדה המקומית ציינה כי..."
-- **מבנה קבוע עם 3 חלקים:**
-  1. "טענות העוררים" — 8-12 סעיפים מקובצים לפי נושא
-  2. "עמדת הוועדה המקומית" — 5-8 סעיפים
-  3. "עמדת מבקשי ההיתר" (אם יש) — 5-10 סעיפים
+- **מבנה: חלק (סעיף עם כותרת-משנה) לכל צד שמופיע בטענות שחולצו למטה** — לפי הכותרות שם (### ...). אל תמציא צדדים שאינם מופיעים ואל תשמיט צד שמופיע.
+  - "טענות העוררים" — 8-12 סעיפים מקובצים לפי נושא.
+  - **צד-משיב/מתנגד: כתוב חלק נפרד לכל כתב-תשובה** שמופיע (למשל "עמדת משיבות 2-3" ו-"עמדת משיבים 4-6" בנפרד) — 5-10 סעיפים לכל אחד. **אל תמזג משיבים שונים לסעיף אחד**, ואם עמדותיהם מנוגדות — שקף זאת.
+  - "עמדת הוועדה המקומית" / "עמדת מבקשי ההיתר" (אם מופיעים) — 5-8 סעיפים.
 - כותרת: "תמצית טענות הצדדים"
 - נאמנות למקור — לא להמציא טענות, אבל כן לאחד ולסכם טענות חוזרות
 - אין ניתוח, אין מסקנות, אין הערכה ("טענה חלשה/חזקה")
 - רק מכתבי טענות מקוריים (לא השלמות טיעון)
 - מספור רציף
-- **יעד אורך: 800-1500 מילים**
+- **יעד אורך: 800-1500 מילים** (יותר כשיש כמה כתבי-תשובה נפרדים — כל צד מקבל את מלוא ההתייחסות)
 
 ## טענות שחולצו (קבץ טענות דומות לנושאים):
 {claims_context}
@@ -272,6 +284,7 @@ BLOCK_PROMPTS = {
 - **ללא כפילות** — הפנה לבלוקים קודמים: "כאמור בסעיף X לעיל"
 - **מספור רציף** — המשך מספור מהבלוק הקודם
 - מותרות כותרות-משנה כשיש נושאים נפרדים לחלוטין
+- **צדדים מרובים** — כשיש כמה משיבים/מתנגדים או כתבי-תשובה נפרדים (ראה הכותרות שתחת "טענות" לעיל, למשל "משיבות 2-3" מול "משיבים 4-6"), התייחס לעמדת כל אחד לגופה ואל תמזג אותם ל"טענות המשיבים" גורפות. אם משיבים שונים נוקטים **עמדות מנוגדות זו לזו** — ציין זאת במפורש והכרע ביניהן בנימוק; אם עמדותיהם משלימות — ניתן לאגד תוך שמירת הייחוס.
 
 ## כיוון מאושר (חובה):
 {direction_context}
@@ -461,17 +474,21 @@ async def write_block(
         if not dir_doc.get("approved"):
             raise ValueError("לא ניתן לכתוב בלוק דיון ללא כיוון מאושר. הפעל brainstorm → approve_direction קודם.")
 
-    # Guard against context overflow before calling claude -p.
-    # Sonnet: 200K context → ~800K chars max; Opus: 200K context → same.
-    # In practice the CLI has crashed on prompts above ~400K chars, so use
-    # that as a conservative ceiling (well below the token limit).
-    _MAX_PROMPT_CHARS = 400_000
+    # Pick the model build by prompt size (#216). Above the 200K-token wall we
+    # escalate to the 1M-context build (`[1m]`) instead of failing the block —
+    # block-yod legitimately carries the whole case as source-context. The 400K
+    # ceiling was an artifact of the old 200K-only build, NOT a model limit.
+    gen_model = GENERATION_MODEL_1M if len(prompt) > _CTX_1M_THRESHOLD_CHARS else GENERATION_MODEL
+
+    # Final guard: even the 1M build is finite (~2M Hebrew chars of input). Cap at
+    # 1.5M chars (~750K tokens) to leave room for output + a safety margin under 1M.
+    _MAX_PROMPT_CHARS = 1_500_000
     if len(prompt) > _MAX_PROMPT_CHARS:
         raise RuntimeError(
             f"Prompt too large for {block_id}: {len(prompt):,} chars "
-            f"(limit {_MAX_PROMPT_CHARS:,}). "
+            f"(limit {_MAX_PROMPT_CHARS:,}, even on the 1M-context build). "
             f"source_context: {len(source_context):,} chars. "
-            f"Reduce documents or call extract_appraiser_facts first."
+            f"Reduce source-context (summaries / appraiser_facts / focused RAG)."
         )
 
     # Call Claude via Claude Code session (no API). #204: pin the model + per-block
@@ -483,14 +500,14 @@ async def write_block(
     content = await claude_session.query(
         prompt,
         timeout=timeout,
-        model=GENERATION_MODEL,
+        model=gen_model,
         effort=effort,
         tools="",  # prose gen — no tool_use → no error_max_turns
     )
 
     sources = await _collect_block_sources(case_id, block_id)
     sources["case_law_ids"] = _precedent_case_law_ids
-    result = _build_result(block_id, content, block_cfg)
+    result = _build_result(block_id, content, block_cfg, model_used=gen_model)
     # Record the EFFECTIVE effort (override wins) so the harness can attribute
     # the measured distance to the effort that actually produced the text.
     if result.get("effort") is not None:
@@ -499,7 +516,8 @@ async def write_block(
     return result
 
 
-def _build_result(block_id: str, content: str, block_cfg: dict) -> dict:
+def _build_result(block_id: str, content: str, block_cfg: dict,
+                  model_used: str = GENERATION_MODEL) -> dict:
     word_count = len(content.split())
     is_ai = block_cfg["model"] == "ai"
     return {
@@ -509,8 +527,9 @@ def _build_result(block_id: str, content: str, block_cfg: dict) -> dict:
         "content": content,
         "word_count": word_count,
         "generation_type": block_cfg["gen_type"],
-        # AI blocks record the pinned model; template blocks record "script".
-        "model_used": GENERATION_MODEL if is_ai else block_cfg["model"],
+        # AI blocks record the model build actually used (200K or [1m], #216);
+        # template blocks record "script".
+        "model_used": model_used if is_ai else block_cfg["model"],
         # The real generation knob (#204). None for template/script blocks.
         "effort": block_cfg.get("effort", DEFAULT_EFFORT) if is_ai else None,
         # DEPRECATED: temperature is not a real knob on Opus 4.7/4.8 (sending it
@@ -611,18 +630,49 @@ async def _build_claims_context(case_id: UUID) -> str:
         # Fallback to all non-block-zayin claims if no source claims exist.
         source_claims = [c for c in claims if c.get("source_document", "") != "block-zayin"] or claims
 
-    lines = []
-    current_role = ""
+    # Group by (party_role, brief) so a multi-litigant side (respondent /
+    # permit_applicant) whose claims come from distinct pleadings reads as
+    # SEPARATE positions — e.g. "כתב תשובה משיבות 2-3" vs "משיבים 4-6" — instead
+    # of one merged "טענות המשיבים" (#224). The brief label is the source
+    # pleading, and which sides split is SPLIT_PARTIES — the SAME rule the
+    # argument aggregator uses, so block ז and the arguments panel agree (G2).
+    from legal_mcp.services.argument_aggregator import SPLIT_PARTIES
+
     role_heb = {"appellant": "טענות העוררים", "respondent": "טענות המשיבים",
                 "committee": "עמדת הוועדה המקומית", "permit_applicant": "עמדת מבקשי ההיתר"}
-    claim_num = 0
+    role_order = ["appellant", "committee", "respondent", "permit_applicant"]
+
+    groups: dict[tuple[str, str], list[dict]] = {}
     for c in source_claims:
-        if c["party_role"] != current_role:
-            current_role = c["party_role"]
-            lines.append(f"\n### {role_heb.get(current_role, current_role)}")
+        role = c.get("party_role", "") or ""
+        # Prefer the stored party_name (stamped by the extractor, #224); fall
+        # back to source_document for legacy claims predating the stamping.
+        if role in SPLIT_PARTIES:
+            brief = (c.get("party_name") or "").strip() or (c.get("source_document") or "").strip()
+        else:
+            brief = ""
+        groups.setdefault((role, brief), []).append(c)
+
+    def _sort_key(k: tuple[str, str]) -> tuple[int, str]:
+        role, brief = k
+        idx = role_order.index(role) if role in role_order else len(role_order)
+        return (idx, brief)
+
+    lines: list[str] = []
+    claim_num = 0
+    for role, brief in sorted(groups.keys(), key=_sort_key):
+        header = role_heb.get(role, role or "טענות נוספות")
+        if brief:
+            header = f"{header} — {brief}"
+        lines.append(f"\n### {header}")
+        for c in groups[(role, brief)]:
             claim_num += 1
             lines.append(f"טענה #{claim_num}: {c['claim_text'][:400]}")
-    lines.append(f"\n**סה\"כ {claim_num} טענות. ענה על כל טענה מהותית; טענות [bundle] — אגד; טענות [skip] — ציון קצר בלבד.**")
+    lines.append(
+        f"\n**סה\"כ {claim_num} טענות. שמור על ההפרדה בין הצדדים לעיל — "
+        f"לכל צד/כתב-תשובה עמדה נפרדת (אל תמזג משיבים שונים). ענה על כל טענה "
+        f"מהותית; טענות [bundle] — אגד; טענות [skip] — ציון קצר בלבד.**"
+    )
     return "\n".join(lines)
 
 

mcp-server/src/legal_mcp/services/case_citation_verification.py

@@ -18,6 +18,7 @@ rows. Reuses the one corpus search + the one authority query + the one radar (G2
 
 from __future__ import annotations
 
+import asyncio
 import logging
 from uuid import UUID
 
@@ -65,25 +66,37 @@ async def build_view(case_number: str) -> dict:
     except Exception as e:  # noqa: BLE001 — radar is best-effort
         logger.warning("citation_verification radar failed for %s: %s", case_number, e)
 
-    out_args: list[dict] = []
-    n_verified = 0
-    for a in args:
-        aid = str(a["id"])
+    async def _fetch(a: dict) -> tuple[list[dict], dict]:
+        """Per-argument corpus search + batched authority — run concurrently.
+
+        Each call is one Voyage embed + one vector search (+ one batched authority
+        query); independent across arguments, so they fan out rather than waterfall.
+        """
         title = (a.get("argument_title") or "").strip()
         topic = (a.get("legal_topic") or "").strip()
         query = f"{ctx} {title}. {topic}".strip()
-
-        hits = []
+        hits: list[dict] = []
         try:
             hits = await precedent_library.search_library(
                 query=query, limit=_SUGGEST_PER_ISSUE, include_halachot=True)
         except Exception as e:  # noqa: BLE001
             logger.warning("citation_verification search failed (%s): %s", title[:30], e)
-
         # Resolve the authority breakdown for the hit set in one batched query.
         clids = [UUID(str(h["case_law_id"])) for h in hits
                  if h.get("case_law_id") and float(h.get("score", 0) or 0) >= _SUGGEST_FLOOR]
         authority = await db.citation_authority(clids) if clids else {}
+        return hits, authority
+
+    # Fan out the expensive per-argument retrieval concurrently — was a sequential
+    # waterfall (N args × Voyage embed + vector search each). gather preserves order.
+    fetched = await asyncio.gather(*(_fetch(a) for a in args)) if args else []
+
+    out_args: list[dict] = []
+    n_verified = 0
+    for a, (hits, authority) in zip(args, fetched):
+        aid = str(a["id"])
+        title = (a.get("argument_title") or "").strip()
+        topic = (a.get("legal_topic") or "").strip()
 
         seen: set[str] = set()
         supporting: list[dict] = []

mcp-server/src/legal_mcp/services/claims_extractor.py

@@ -368,8 +368,15 @@ async def extract_and_store_claims(
 
     # Determine claim_type from document type and title
     claim_type = _infer_claim_type(doc_type, source_name)
+    # Stamp party_name at the source (#224). For a multi-litigant side
+    # (respondent / permit_applicant) the brief label IS the source pleading, so
+    # opposing briefs (משיבות 2-3 vs משיבים 4-6) stay distinct downstream without
+    # the aggregator/block-writer having to re-derive it. SPLIT_PARTIES is the
+    # single rule shared with the aggregator (G2); single-voice sides stay ''.
+    from legal_mcp.services.argument_aggregator import SPLIT_PARTIES
     for c in claims:
         c["claim_type"] = claim_type
+        c["party_name"] = source_name if c.get("party_role") in SPLIT_PARTIES else ""
 
     stored = await db.store_claims(case_id, claims, source_document=source_name)
     # Mark this document analysed (WS2 / #201). store_claims already replaced

mcp-server/src/legal_mcp/services/claude_session.py

@@ -146,8 +146,15 @@ async def query(
     """
     full_prompt = f"{system}\n\n{prompt}" if system else prompt
 
-    if len(full_prompt) > 150_000:
-        logger.warning("Large prompt: %d chars — may hit context limits", len(full_prompt))
+    # ~350K chars ≈ the 200K-token wall of the default Opus build (~2 Hebrew
+    # chars/token). Above it callers should pass the 1M-context model id
+    # (`claude-opus-4-8[1m]`) — block_writer escalates automatically (#216).
+    if len(full_prompt) > 350_000 and "[1m]" not in (model or ""):
+        logger.warning(
+            "Large prompt: %d chars on a non-[1m] model (%s) — may overflow the "
+            "200K-token context. Pass claude-opus-4-8[1m] for the 1M window.",
+            len(full_prompt), model or "CLI-default",
+        )
 
     cmd = [
         "claude", "-p",

mcp-server/src/legal_mcp/services/db.py

@@ -1888,6 +1888,50 @@ CREATE INDEX IF NOT EXISTS idx_documents_claims_pending
 """
 
 
+# V50 (#224): per-respondent-brief separation. The aggregator groups respondent /
+# permit_applicant claims by their source pleading — each joint response brief
+# (e.g. "כתב תשובה משיבות 2-3" vs "משיבים 4-6") is one coherent litigation
+# position, so opposing briefs no longer collapse into a single "respondent"
+# bucket. ``party_name`` carries the brief label; appellant/committee stay ''.
+SCHEMA_V50_SQL = """
+ALTER TABLE legal_arguments ADD COLUMN IF NOT EXISTS party_name TEXT NOT NULL DEFAULT '';
+CREATE INDEX IF NOT EXISTS idx_legal_arguments_party_name
+    ON legal_arguments(case_id, party, party_name);
+"""
+
+
+# V51 (#226): hearing-attendance provenance for the "מה קרה בדיון" panel. The
+# ערר-hearing protocol names who actually appeared (עוררים/משיבים + their
+# counsel) and, when stated, the presiding panel. ``_extract_header`` already
+# extracts this feed but only ``hearing_date`` had a canonical home (on cases);
+# the attendee lists were returned as provenance and dropped. This column gives
+# them a home so the panel can display them.
+#
+# This is DISTINCT from ``decisions.panel_members`` (G2, not a parallel path):
+# that column is the authoring tribunal recorded on a specific written-decision
+# version (the DOCX signature block). ``hearing_attendees`` is a snapshot of who
+# was present at the *hearing*, extracted from the protocol — hearing-event
+# provenance, not decision-authorship. Shape:
+#   {"panel_members": [...], "appellants_present": [...], "respondents_present": [...]}
+SCHEMA_V51_SQL = """
+ALTER TABLE cases ADD COLUMN IF NOT EXISTS hearing_attendees JSONB NOT NULL DEFAULT '{}';
+"""
+
+
+# V52 (#224): backfill claims.party_name for existing rows. The extractor now
+# stamps party_name = source_document for the multi-litigant sides at write time
+# (claims_extractor), but claims stored before that shipped have party_name=''.
+# For those, the brief label is the source pleading — same rule the aggregator
+# uses (respondent / permit_applicant split). Idempotent: only fills empties, so
+# re-running is a no-op. Single-voice sides (appellant/committee) stay ''.
+SCHEMA_V52_SQL = """
+UPDATE claims SET party_name = source_document
+WHERE party_role IN ('respondent', 'permit_applicant')
+  AND COALESCE(party_name, '') = ''
+  AND COALESCE(source_document, '') <> '';
+"""
+
+
 # Stable, arbitrary key for the session-level advisory lock that serialises
 # schema DDL across processes. Every short-lived process (cron drains, services)
 # re-runs the idempotent migrations on startup; without this lock two processes
@@ -1959,6 +2003,9 @@ async def _apply_schema_ddl(conn: asyncpg.Connection) -> None:
         await conn.execute(SCHEMA_V47_SQL)
         await conn.execute(SCHEMA_V48_SQL)
         await conn.execute(SCHEMA_V49_SQL)
+        await conn.execute(SCHEMA_V50_SQL)
+        await conn.execute(SCHEMA_V51_SQL)
+        await conn.execute(SCHEMA_V52_SQL)
 
 
 async def init_schema() -> None:
@@ -2181,7 +2228,7 @@ async def update_case(case_id: UUID, **fields) -> dict | None:
     set_clauses = []
     values = []
     for i, (key, val) in enumerate(fields.items(), start=2):
-        if key in ("appellants", "respondents", "tags"):
+        if key in ("appellants", "respondents", "tags", "hearing_attendees"):
             val = json.dumps(val)
         set_clauses.append(f"{key} = ${i}")
         values.append(val)
@@ -2194,7 +2241,7 @@ async def update_case(case_id: UUID, **fields) -> dict | None:
 
 def _row_to_case(row: asyncpg.Record) -> dict:
     d = dict(row)
-    for field in ("appellants", "respondents", "tags"):
+    for field in ("appellants", "respondents", "tags", "hearing_attendees"):
         if isinstance(d.get(field), str):
             d[field] = json.loads(d[field])
     d["id"] = str(d["id"])

mcp-server/src/legal_mcp/services/docx_exporter.py

@@ -43,10 +43,16 @@ HEBREW_FONT = "David"
 
 
 def _mark_run_rtl(run) -> None:
-    """Force David font on all four slots, then add <w:rtl/>."""
+    """Force David font on all four slots, then add <w:rtl/>.
+
+    INV-EX4 note: this reinforces the template's OWN font (David, == the
+    template's cs slot) to work around Word's RTL ascii-slot fallback to
+    Times New Roman — it is NOT an ad-hoc font/size override. Marked
+    INV-EX4-ok for the single-source guard for that reason.
+    """
     rPr = run._r.get_or_add_rPr()
     if rPr.find(qn("w:rFonts")) is None:
-        fonts = OxmlElement("w:rFonts")
+        fonts = OxmlElement("w:rFonts")  # INV-EX4-ok: reinforces template font David
         fonts.set(qn("w:ascii"), HEBREW_FONT)
         fonts.set(qn("w:hAnsi"), HEBREW_FONT)
         fonts.set(qn("w:cs"), HEBREW_FONT)
@@ -393,8 +399,15 @@ _INTERIM_BLOCK_ORDER = [
 ]
 
 
-def _draft_filename_prefix(mode: str) -> str:
-    return "טיוטת-ביניים" if mode == "interim" else "טיוטה"
+def _draft_naming(mode: str) -> tuple[str, str]:
+    """(filename prefix, version separator) per export mode.
+
+    interim → ``טיוטה-טענות_הצדדים_{N}.docx`` (chair-requested naming);
+    final   → ``טיוטה-v{N}.docx``.
+    """
+    if mode == "interim":
+        return "טיוטה-טענות_הצדדים", "_"
+    return "טיוטה", "-v"
 
 
 async def export_decision(
@@ -479,16 +492,19 @@ async def export_decision(
     if not output_path:
         export_dir = config.find_case_dir(case["case_number"]) / "exports"
         export_dir.mkdir(parents=True, exist_ok=True)
-        prefix = _draft_filename_prefix(mode)
-        existing = sorted(export_dir.glob(f"{prefix}-v*.docx"))
+        prefix, sep = _draft_naming(mode)
+        existing = sorted(export_dir.glob(f"{prefix}{sep}*.docx"))
         next_ver = 1
         for p in existing:
             try:
-                ver = int(p.stem.split("-v")[1])
+                # Version is the trailing integer after the separator. Using
+                # rsplit keeps this correct even when the prefix itself contains
+                # the separator char (e.g. "טיוטה-טענות_הצדדים" with sep="_").
+                ver = int(p.stem.rsplit(sep, 1)[1])
                 next_ver = max(next_ver, ver + 1)
             except (IndexError, ValueError):
                 pass
-        output_path = str(export_dir / f"{prefix}-v{next_ver}.docx")
+        output_path = str(export_dir / f"{prefix}{sep}{next_ver}.docx")
 
     # Persist through the storage layer (INV-STG1). Under the filesystem
     # backend the bytes land at output_path exactly as before; a caller-

mcp-server/src/legal_mcp/services/docx_reviser.py

@@ -199,27 +199,86 @@ def _find_enclosing_paragraph(element: etree._Element) -> etree._Element | None:
     return None
 
 
+# ── Template style resolution (INV-EX4) ────────────────────────────
+# Every inserted paragraph carries a <w:pStyle> pointing at one of the
+# template's named styles — font (David), size, and indent come from
+# styles.xml, NOT from hand-set rFonts/sz here. Single template = single
+# style source (INV-EX4 / G2). docx_exporter.py applies the same styles by
+# name via python-docx; here (raw lxml) we resolve name → styleId.
+
+# StyleType → candidate style names in the template (case-insensitive).
+# Mirrors the names docx_exporter uses (paragraph.style = "Heading 2"/"Normal"/"Quote").
+_STYLE_NAME_CANDIDATES: dict[str, tuple[str, ...]] = {
+    "body": ("Normal", "Default Paragraph Font", "Body Text"),
+    "bold": ("Normal", "Body Text"),
+    "heading": ("Heading 2", "heading 2"),
+    "quote": ("Quote", "quote", "Intense Quote"),
+}
+
+
+def _resolve_style_ids(members: dict[str, bytes]) -> dict[str, str]:
+    """Map each StyleType → the template's w:styleId, read from styles.xml.
+
+    Returns {} when styles.xml is absent — callers then fall back to
+    minimal inline emphasis (so a missing style never crashes rendering).
+    """
+    raw = members.get("word/styles.xml")
+    if not raw:
+        return {}
+    try:
+        styles_tree = etree.fromstring(raw)
+    except etree.XMLSyntaxError:  # pragma: no cover - corrupt template
+        return {}
+    # Build a lowercase lookup over both the style's display name and its id.
+    by_key: dict[str, str] = {}
+    for style in styles_tree.iterfind(".//w:style", NSMAP):
+        sid = style.get(_w("styleId"))
+        if not sid:
+            continue
+        by_key.setdefault(sid.lower(), sid)
+        name_el = style.find(_w("name"))
+        if name_el is not None:
+            nm = name_el.get(_w("val"))
+            if nm:
+                by_key.setdefault(nm.lower(), sid)
+    resolved: dict[str, str] = {}
+    for style_type, candidates in _STYLE_NAME_CANDIDATES.items():
+        for cand in candidates:
+            sid = by_key.get(cand.lower())
+            if sid is not None:
+                resolved[style_type] = sid
+                break
+    return resolved
+
+
+def _emphasis_fallback(style: StyleType, *, has_pstyle: bool) -> tuple[bool, bool]:
+    """(bold, italic) for a run. When the paragraph carries a resolved
+    template pStyle, emphasis comes from the style (return no inline override),
+    except 'bold' which is inline emphasis with no dedicated style. When no
+    pStyle resolved, fall back to minimal inline emphasis so heading/quote
+    stay visually distinct."""
+    if style == "bold":
+        return True, False
+    if has_pstyle:
+        return False, False
+    return (style == "heading"), (style == "quote")
+
+
 # ── Paragraph builders ─────────────────────────────────────────────
 
 
-def _build_run(text: str, *, bold: bool = False, italic: bool = False,
-               font: str = "David", size_half_pt: int | None = None) -> etree._Element:
-    """Build a w:r (run) element with RTL/David defaults and given text."""
+def _build_run(text: str, *, bold: bool = False, italic: bool = False) -> etree._Element:
+    """Build a w:r (run): RTL marker + optional bold/italic + text.
+
+    Font (David) and size are NOT set here — they derive from the paragraph's
+    template style (INV-EX4). The <w:rtl/> marker is still required so Word
+    picks the style's complex-script (cs=David) slot rather than the ascii
+    (Times New Roman) slot for Hebrew (same reason as _mark_run_rtl in
+    analysis_docx_exporter.py).
+    """
     r = etree.Element(_w("r"))
     rPr = etree.SubElement(r, _w("rPr"))
 
-    rFonts = etree.SubElement(rPr, _w("rFonts"))
-    rFonts.set(_w("ascii"), font)
-    rFonts.set(_w("hAnsi"), font)
-    rFonts.set(_w("cs"), font)
-    rFonts.set(_w("hint"), "cs")
-
-    if size_half_pt is not None:
-        sz = etree.SubElement(rPr, _w("sz"))
-        sz.set(_w("val"), str(size_half_pt))
-        szCs = etree.SubElement(rPr, _w("szCs"))
-        szCs.set(_w("val"), str(size_half_pt))
-
     if bold:
         etree.SubElement(rPr, _w("b"))
         etree.SubElement(rPr, _w("bCs"))
@@ -235,28 +294,36 @@ def _build_run(text: str, *, bold: bool = False, italic: bool = False,
     return r
 
 
-def _build_paragraph(text: str, *, style: StyleType = "body") -> etree._Element:
-    """Build a w:p (paragraph) with RTL + David + given text."""
+def _build_paragraph(text: str, *, style: StyleType = "body",
+                     style_id: str | None = None) -> etree._Element:
+    """Build a w:p (paragraph) carrying the template's named style.
+
+    The paragraph references `style_id` (resolved from the template's
+    styles.xml) via <w:pStyle> — font/size/indent come from there, not from
+    hand-set rFonts/sz (INV-EX4). The only explicit paragraph-level override
+    is the canonical body alignment jc=both, matching docx_exporter's
+    _set_paragraph_jc (which forces both over style-center headings).
+    """
     p = etree.Element(_w("p"))
     pPr = etree.SubElement(p, _w("pPr"))
+
+    if style_id:
+        pStyle = etree.SubElement(pPr, _w("pStyle"))
+        pStyle.set(_w("val"), style_id)
+
     bidi = etree.SubElement(pPr, _w("bidi"))
     bidi.set(_w("val"), "1")
 
-    # Right alignment for body/RTL
+    # Canonical body alignment (two-sided justification) — overrides any
+    # style-center, exactly as docx_exporter._set_paragraph_jc does.
     jc = etree.SubElement(pPr, _w("jc"))
-    jc.set(_w("val"), "right")
+    jc.set(_w("val"), "both")
 
     rPr_p = etree.SubElement(pPr, _w("rPr"))
     etree.SubElement(rPr_p, _w("rtl"))
 
-    bold = style in ("heading", "bold")
-    italic = style == "quote"
-    size = None
-    if style == "heading":
-        size = 28  # 14pt
-    elif style == "quote":
-        size = 22  # 11pt
-    run = _build_run(text, bold=bold, italic=italic, size_half_pt=size)
+    bold, italic = _emphasis_fallback(style, has_pstyle=bool(style_id))
+    run = _build_run(text, bold=bold, italic=italic)
     p.append(run)
     return p
 
@@ -275,7 +342,7 @@ def _wrap_in_ins(elements: list[etree._Element], *, ins_id: int,
 
 def _make_tracked_paragraph_insert(
     text: str, *, style: StyleType, ins_id: int, author: str, date_iso: str,
-    mark_id: int | None = None,
+    mark_id: int | None = None, style_id: str | None = None,
 ) -> etree._Element:
     """Build a whole tracked-inserted paragraph.
 
@@ -286,7 +353,7 @@ def _make_tracked_paragraph_insert(
     """
     if mark_id is None:
         mark_id = ins_id
-    p = _build_paragraph(text, style=style)
+    p = _build_paragraph(text, style=style, style_id=style_id)
     pPr = p.find(_w("pPr"))
     assert pPr is not None
     rPr = pPr.find(_w("rPr"))
@@ -339,6 +406,7 @@ def _apply_insert(
     ins_id: int,
     author: str,
     date_iso: str,
+    style_ids: dict[str, str] | None = None,
 ) -> RevisionResult:
     """Apply insert_after / insert_before relative to a bookmark."""
     start, end = _find_bookmark(document_tree, revision.anchor_bookmark)
@@ -362,6 +430,7 @@ def _apply_insert(
         revision.content, style=revision.style,
         ins_id=ins_id, mark_id=ins_id + 1,
         author=author, date_iso=date_iso,
+        style_id=(style_ids or {}).get(revision.style),
     )
 
     parent = enclosing_p.getparent()
@@ -407,6 +476,7 @@ def _apply_replace(
     del_id: int,
     author: str,
     date_iso: str,
+    style_ids: dict[str, str] | None = None,
 ) -> RevisionResult:
     """Replace = delete the existing paragraph + insert new one after it."""
     start, end = _find_bookmark(document_tree, revision.anchor_bookmark)
@@ -428,6 +498,7 @@ def _apply_replace(
         revision.content, style=revision.style,
         ins_id=ins_id, mark_id=ins_id + 1,
         author=author, date_iso=date_iso,
+        style_id=(style_ids or {}).get(revision.style),
     )
     idx = list(parent).index(enclosing_p)
     parent.insert(idx + 1, new_p)
@@ -473,6 +544,10 @@ def apply_tracked_revisions(
     members, doc_tree, settings_tree = _load_docx_xml(source_path)
     _ensure_track_revisions(settings_tree)
 
+    # Resolve template style ids once (INV-EX4) — inserted paragraphs inherit
+    # font/size/indent from these named styles instead of hand-set rFonts/sz.
+    style_ids = _resolve_style_ids(members)
+
     next_id = _next_revision_id(doc_tree)
 
     batch = RevisionBatchResult()
@@ -480,7 +555,8 @@ def apply_tracked_revisions(
         try:
             if rev.type in ("insert_after", "insert_before"):
                 result = _apply_insert(doc_tree, rev, ins_id=next_id,
-                                       author=author, date_iso=date_iso)
+                                       author=author, date_iso=date_iso,
+                                       style_ids=style_ids)
                 # insert consumes 2 IDs: run-wrapper + paragraph-mark
                 next_id += 2
             elif rev.type == "delete":
@@ -490,7 +566,8 @@ def apply_tracked_revisions(
             elif rev.type == "replace":
                 result = _apply_replace(doc_tree, rev,
                                         ins_id=next_id, del_id=next_id + 2,
-                                        author=author, date_iso=date_iso)
+                                        author=author, date_iso=date_iso,
+                                        style_ids=style_ids)
                 # replace consumes 3 IDs: ins-run, ins-mark, del
                 next_id += 3
             else:

mcp-server/src/legal_mcp/services/protocol_analyzer.py

@@ -120,15 +120,47 @@ HEADER_PROMPT = """אתה מחלץ נתוני-כותרת מפרוטוקול די
 MAX_PROTOCOL_CHARS = 120_000
 
 
-def _find_protocol(docs: list[dict]) -> dict | None:
-    """The protocol document for the case, if present (doc_type or title)."""
-    for d in docs:
-        if d.get("doc_type") == "protocol":
-            return d
-    for d in docs:
-        if "פרוטוקול" in (d.get("title") or ""):
-            return d
+def _protocol_scope(doc: dict) -> str:
+    """Protocol scope from ``metadata.protocol_scope``.
+
+    Absent/'' defaults to ``'appeal'`` (ועדת הערר), matching the
+    document-type-editor convention where only the non-default ``'lower'``
+    (ועדה מקומית/מחוזית) is persisted.
+    """
+    meta = doc.get("metadata") or {}
+    return (meta.get("protocol_scope") or "").strip() or "appeal"
+
+
+def _find_protocol(
+    docs: list[dict], document_id: UUID | None = None,
+) -> dict | None:
+    """The ועדת-הערר hearing protocol to compare against the pleadings.
+
+    - ``document_id`` given → return exactly that document (explicit target),
+      so a case with several protocols can be pointed at the right hearing.
+    - Otherwise prefer a protocol scoped to the appeals committee
+      (``scope != 'lower'``). A ``'lower'`` protocol is the local/district
+      committee's proceedings — it feeds background (block ו) only and must NOT
+      drive the hearing-vs-pleadings comparison (#223). If every protocol is
+      lower-scoped there is no ערר-hearing to compare → return None.
+    - Among appeal-scoped protocols the most recent (by ``created_at``) wins —
+      the operative hearing; ties fall back to list order.
+    """
+    if document_id is not None:
+        return next(
+            (d for d in docs if str(d.get("id")) == str(document_id)), None,
+        )
+
+    protocols = [d for d in docs if d.get("doc_type") == "protocol"]
+    if not protocols:
+        # Legacy fallback: untyped docs whose title says "פרוטוקול".
+        protocols = [d for d in docs if "פרוטוקול" in (d.get("title") or "")]
+
+    appeal_scoped = [d for d in protocols if _protocol_scope(d) != "lower"]
+    if not appeal_scoped:
         return None
+    appeal_scoped.sort(key=lambda d: d.get("created_at") or "", reverse=True)
+    return appeal_scoped[0]
 
 
 def _compact_arguments(arguments: list[dict]) -> list[dict]:
@@ -227,21 +259,39 @@ async def _extract_header(protocol_text: str, case_id: UUID) -> dict:
         except ValueError:
             logger.info("protocol header: unparseable hearing_date %r", hearing_date)
 
+    # Attendees → cases.hearing_attendees (#226). Unlike hearing_date there is no
+    # chair-entry path to protect, and re-running to point at the correct ערר
+    # protocol (#223) must refresh who appeared — so write whenever the operative
+    # protocol yielded any names, but never clobber good data with an empty
+    # extraction (all-empty feed → leave the prior snapshot intact).
+    attendees = {
+        "panel_members": feed.get("panel_members") or [],
+        "appellants_present": feed.get("appellants_present") or [],
+        "respondents_present": feed.get("respondents_present") or [],
+    }
+    if any(attendees.values()):
+        await db.update_case(case_id, hearing_attendees=attendees)
+        applied["hearing_attendees"] = True
+
     return {
         "status": "ok",
         "feed": {
             "hearing_date": hearing_date,
-            "panel_members": feed.get("panel_members") or [],
-            "appellants_present": feed.get("appellants_present") or [],
-            "respondents_present": feed.get("respondents_present") or [],
+            **attendees,
         },
         "applied_to_case": applied,
     }
 
 
-async def analyze_protocol(case_id: UUID) -> dict:
+async def analyze_protocol(
+    case_id: UUID, target_document_id: UUID | None = None,
+) -> dict:
     """Comparative analysis of the case's hearing protocol vs. its pleadings.
 
+    ``target_document_id`` pins the analysis to a specific protocol document —
+    required when a case holds several protocols (e.g. a lower-committee annex
+    plus the ועדת-הערר hearing) and the auto-pick would be ambiguous (#223).
+
     1. Locates the protocol document.
     2. Pulls the aggregated legal_arguments (the written-pleadings baseline).
     3. Asks Claude (Opus 4.8, effort=high) to classify each as dropped /
@@ -252,11 +302,20 @@ async def analyze_protocol(case_id: UUID) -> dict:
     Returns a serializable summary dict.
     """
     docs = await db.list_documents(case_id)
-    protocol = _find_protocol(docs)
+    protocol = _find_protocol(docs, document_id=target_document_id)
     if not protocol:
+        if target_document_id is not None:
             return {
                 "status": "no_protocol",
-            "message": "לא נמצא פרוטוקול דיון בתיק (doc_type='protocol'). העלה פרוטוקול והרץ שוב.",
+                "message": f"מסמך {target_document_id} לא נמצא בתיק.",
+            }
+        return {
+            "status": "no_protocol",
+            "message": (
+                "לא נמצא פרוטוקול ועדת-ערר בתיק (doc_type='protocol' עם "
+                "protocol_scope שאינו 'lower'). העלה פרוטוקול-דיון, או תייג את "
+                "הפרוטוקול הקיים כ-appeal, והרץ שוב."
+            ),
         }
 
     document_id = UUID(protocol["id"])

mcp-server/src/legal_mcp/services/research_md.py

@@ -15,8 +15,14 @@ import os
 import re
 from datetime import datetime
 from pathlib import Path
+from zoneinfo import ZoneInfo
 from typing import Any
 
+# Display timezone (INV-UI9): timestamps surfaced to humans are rendered in
+# Israel time deterministically, regardless of the process clock (the container
+# may run UTC or Asia/Jerusalem). Storage stays UTC; this is display-only.
+IL_TZ = ZoneInfo("Asia/Jerusalem")
+
 # Placeholder strings — any of these means "not yet filled"
 CHAIR_POSITION_PLACEHOLDERS = (
     "[ימולא ע\"י יו\"ר הוועדה]",
@@ -191,7 +197,7 @@ def parse(file_path: Path) -> dict[str, Any]:
     date_str = date_match.group(1) if date_match else ""
 
     stat = file_path.stat()
-    mtime_iso = datetime.fromtimestamp(stat.st_mtime).isoformat()
+    mtime_iso = datetime.fromtimestamp(stat.st_mtime, tz=IL_TZ).isoformat()
 
     result: dict[str, Any] = {
         "header": {
@@ -354,7 +360,7 @@ def update_chair_position(
         "saved": True,
         "section_id": section_id,
         "preview": preview,
-        "timestamp": datetime.now().isoformat(),
+        "timestamp": datetime.now(IL_TZ).isoformat(),
     }
 
 

mcp-server/src/legal_mcp/tools/cases.py

@@ -13,6 +13,7 @@ from uuid import UUID
 import httpx
 
 from legal_mcp import config
+from legal_mcp.case_status_model import STATUS_ORDER  # status SSoT
 from legal_mcp.services import audit, db, extractor, git_sync, practice_area as pa
 from legal_mcp.tools.envelope import empty, err, ok  # GAP-48: SSoT envelope
 
@@ -341,16 +342,8 @@ async def case_update(
     """
     from datetime import date as date_type
 
-    # Ordered core lifecycle — regression protection (forward-only).
-    # Single source of truth, mirrored by web-ui/src/lib/api/case-status.ts and
-    # models.CaseStatus. Trimmed from 17 → 10 (decorative statuses removed).
-    STATUS_ORDER = [
-        "new", "processing", "documents_ready",
-        "outcome_set", "direction_approved",
-        "qa_review", "drafted",
-        "exported", "reviewed", "final",
-    ]
-
+    # Ordered core lifecycle (forward-only regression guard). Single source of
+    # truth: legal_mcp/case_status_model.py — STATUS_ORDER/BY_KEY derive from it.
     case = await db.get_case_by_number(case_number)
     if not case:
         return err(f"תיק {case_number} לא נמצא.")
@@ -363,6 +356,9 @@ async def case_update(
         # Only update if advancing or status is unknown to the order
         if new_idx >= cur_idx or new_idx == -1:
             fields["status"] = status
+            # per-status on_enter hook (case_status_model) — the single place a
+            # status's behaviour is declared. No dispatcher wired yet; when the
+            # first action is added, dispatch BY_KEY[status].on_enter here.
     if title:
         fields["title"] = title
     if subject:

mcp-server/src/legal_mcp/tools/documents.py

@@ -433,17 +433,24 @@ ALLOWED_DOC_TYPES = {
 # Allowed appraiser_side values; '' (empty) clears the tag.
 ALLOWED_APPRAISER_SIDES = {"committee", "appellant", "deciding", ""}
 
+# Allowed protocol_scope values (protocol docs only); mirrors web/app.py.
+# appeal = ועדת הערר (default when absent) · lower = ועדה מקומית/מחוזית.
+ALLOWED_PROTOCOL_SCOPES = {"appeal", "lower"}
+
 
 async def document_update(
     case_number: str,
     doc_id: str,
     doc_type: str = "",
     appraiser_side: str = "",
+    is_post_hearing: bool | None = None,
+    protocol_scope: str = "",
 ) -> str:
-    """עדכון תיוג מסמך — doc_type ו/או appraiser_side. ריק = אין שינוי.
+    """עדכון תיוג מסמך — doc_type / appraiser_side / is_post_hearing / protocol_scope.
 
-    הולידציה זהה ל-PATCH endpoint ב-web/app.py. appraiser_side נשמר ב-
-    documents.metadata JSONB (מתפרסם משם ע"י extract_appraiser_facts).
+    הולידציה זהה ל-PATCH endpoint ב-web/app.py. הדגלים נשמרים ב-
+    documents.metadata JSONB (appraiser_side מתפרסם משם ע"י extract_appraiser_facts;
+    is_post_hearing נצרך ע"י כותב בלוק-ח; protocol_scope מסווג פרוטוקול בטאב-הסקירה).
 
     Args:
         case_number: מספר תיק הערר (לאישור שייכות)
@@ -452,6 +459,9 @@ async def document_update(
                   permit/appraisal/exhibit/objection/reference). ריק = אין שינוי.
         appraiser_side: ערך חדש (committee/appellant/deciding). ריק = אין שינוי;
                         העבר במפורש מחרוזת ריקה לא-default אם רוצים לנקות.
+        is_post_hearing: True = התקבל אחרי הדיון · False = מנקה את הדגל · None = אין שינוי.
+        protocol_scope: היקף פרוטוקול (appeal=ועדת הערר / lower=ועדה מקומית-מחוזית).
+                        ריק = אין שינוי. רלוונטי רק כשהסוג protocol.
     """
     case = await db.get_case_by_number(case_number)
     if not case:
@@ -477,17 +487,37 @@ async def document_update(
                        data={"allowed": sorted(ALLOWED_DOC_TYPES)})
         updates["doc_type"] = doc_type
 
-    # appraiser_side is optional. The MCP tool can't distinguish "skip" from
-    # "set to empty string", so we use the convention: only update if non-empty.
-    # To clear, the operator must edit metadata directly (rare).
+    # appraiser_side / protocol_scope are optional. The MCP tool can't
+    # distinguish "skip" from "set to empty string", so we use the convention:
+    # only update if non-empty. To clear, edit metadata directly (rare).
+    # is_post_hearing is a tri-state bool: None skips, True/False set/clear.
+    metadata = doc.get("metadata") or {}
+    if isinstance(metadata, str):
+        metadata = json.loads(metadata)
+    metadata_dirty = False
+
     if appraiser_side:
         if appraiser_side not in ALLOWED_APPRAISER_SIDES:
             return err(f"appraiser_side לא תקין: {appraiser_side}",
                        data={"allowed": sorted(s for s in ALLOWED_APPRAISER_SIDES if s)})
-        metadata = doc.get("metadata") or {}
-        if isinstance(metadata, str):
-            metadata = json.loads(metadata)
         metadata["appraiser_side"] = appraiser_side
+        metadata_dirty = True
+
+    if is_post_hearing is not None:
+        if is_post_hearing:
+            metadata["is_post_hearing"] = True
+        else:
+            metadata.pop("is_post_hearing", None)
+        metadata_dirty = True
+
+    if protocol_scope:
+        if protocol_scope not in ALLOWED_PROTOCOL_SCOPES:
+            return err(f"protocol_scope לא תקין: {protocol_scope}",
+                       data={"allowed": sorted(ALLOWED_PROTOCOL_SCOPES)})
+        metadata["protocol_scope"] = protocol_scope
+        metadata_dirty = True
+
+    if metadata_dirty:
         updates["metadata"] = metadata
 
     if not updates:

mcp-server/src/legal_mcp/tools/drafting.py

@@ -580,7 +580,7 @@ async def get_appraiser_facts(case_number: str) -> str:
         return err(str(e))
 
 
-async def analyze_protocol(case_number: str) -> str:
+async def analyze_protocol(case_number: str, document_id: str = "") -> str:
     """ניתוח השוואתי של פרוטוקול-דיון מול כתבי-הטענות (WS4 / #203).
 
     מזהה אילו טענות **ירדו** (נזנחו בדיון), אילו **חוזקו**, ואילו **עלו חדשות**,
@@ -588,12 +588,17 @@ async def analyze_protocol(case_number: str) -> str:
     ל"ידע-התיק" (טבלת protocol_analysis), זמינה לסוכני הניתוח והכתיבה; תאריך-הדיון
     מוזן חזרה לעמודה הקנונית cases.hearing_date.
 
+    בוחר אוטומטית את פרוטוקול **ועדת-הערר** (protocol_scope שאינו 'lower');
+    פרוטוקול ועדה-מקומית ('lower') אינו נכנס להשוואה. כשיש כמה פרוטוקולי-ערר —
+    ציין document_id כדי לכוון לדיון המדויק (#223).
+
     דורש פרוטוקול מתויג doc_type='protocol' + טיעונים מאוגדים
     (aggregate_claims_to_arguments). רץ עם Claude מקומי (Opus 4.8, effort=high);
     re-run מחליף את הניתוח הקודם לאותו פרוטוקול (idempotent).
 
     Args:
         case_number: מספר תיק הערר
+        document_id: מזהה מסמך-הפרוטוקול לכיוון מדויק (ריק = בחירה אוטומטית)
     """
     from legal_mcp.services import protocol_analyzer
 
@@ -601,8 +606,16 @@ async def analyze_protocol(case_number: str) -> str:
     if not case:
         return err(f"תיק {case_number} לא נמצא.")
     case_id = UUID(case["id"])
+    target_doc_id: UUID | None = None
+    if document_id.strip():
         try:
-        result = await protocol_analyzer.analyze_protocol(case_id)
+            target_doc_id = UUID(document_id.strip())
+        except ValueError:
+            return err(f"document_id לא תקין: {document_id}")
+    try:
+        result = await protocol_analyzer.analyze_protocol(
+            case_id, target_document_id=target_doc_id,
+        )
         await audit.log_action_safe(
             "analyze_protocol", case_id=case_id,
             details={"status": result.get("status"), "total": result.get("total", 0)},
@@ -703,7 +716,7 @@ async def write_interim_draft(case_number: str, instructions: str = "") -> str:
 async def export_interim_draft(case_number: str, output_path: str = "") -> str:
     """ייצוא טיוטת ביניים ל-DOCX — אותו עיצוב של טיוטה רגילה (David, RTL,
     bookmarks), אבל בסדר חדש: רקע → תכניות+היתרים → טענות → הליכים, ללא
-    דיון/סיכום/חתימות. שם הקובץ: טיוטת-ביניים-v{N}.docx.
+    דיון/סיכום/חתימות. שם הקובץ: טיוטה-טענות_הצדדים_{N}.docx.
 
     Args:
         case_number: מספר תיק הערר

mcp-server/tests/test_docx_reviser.py

@@ -130,6 +130,16 @@ def _find_ins_with_runs(tree: etree._Element) -> etree._Element | None:
     return None
 
 
+def _find_enclosing_p(element: etree._Element) -> etree._Element | None:
+    """Walk up to the enclosing <w:p>."""
+    cur = element
+    while cur is not None:
+        if cur.tag == _w("p"):
+            return cur
+        cur = cur.getparent()
+    return None
+
+
 def test_insert_after_ins_has_author_and_date(sample_docx: Path, tmp_path: Path) -> None:
     out = tmp_path / "out.docx"
     rev = Revision(id="r1", type="insert_after",
@@ -147,7 +157,10 @@ def test_insert_after_ins_has_author_and_date(sample_docx: Path, tmp_path: Path)
     assert date_str.endswith("Z")  # ISO 8601 UTC
 
 
-def test_insert_after_uses_rtl_and_david(sample_docx: Path, tmp_path: Path) -> None:
+def test_insert_after_uses_rtl_and_template_style(sample_docx: Path, tmp_path: Path) -> None:
+    """INV-EX4: an inserted paragraph carries the template's named style via
+    <w:pStyle> and an <w:rtl/> run marker — but does NOT hand-set rFonts/sz
+    (font/size come from the style). jc is the canonical body 'both'."""
     out = tmp_path / "out.docx"
     rev = Revision(id="r1", type="insert_after",
                    anchor_bookmark="block-alef", content="מוסף")
@@ -162,10 +175,22 @@ def test_insert_after_uses_rtl_and_david(sample_docx: Path, tmp_path: Path) -> N
     assert run is not None
     rPr = run.find(_w("rPr"))
     assert rPr is not None
+    # RTL marker preserved (so Word picks the style's cs=David slot)…
     assert rPr.find(_w("rtl")) is not None
-    rFonts = rPr.find(_w("rFonts"))
-    assert rFonts is not None
-    assert rFonts.get(_w("ascii")) == "David"
+    # …but NO hand-set font/size on the run (INV-EX4).
+    assert rPr.find(_w("rFonts")) is None
+    assert rPr.find(_w("sz")) is None
+    assert rPr.find(_w("szCs")) is None
+
+    # The enclosing paragraph references a template style + canonical jc=both.
+    para = _find_enclosing_p(ins)
+    assert para is not None
+    pPr = para.find(_w("pPr"))
+    assert pPr is not None
+    pStyle = pPr.find(_w("pStyle"))
+    assert pStyle is not None and pStyle.get(_w("val"))  # Normal → real styleId
+    jc = pPr.find(_w("jc"))
+    assert jc is not None and jc.get(_w("val")) == "both"
 
 
 # ── apply_tracked_revisions: insert_before ────────────────────────

mcp-server/tests/test_docx_template_single_source.py

@@ -0,0 +1,76 @@
+"""שומר INV-EX4 — מקור-תבנית-יחיד לכל מפיקי-מסמכי-ההחלטה.
+
+כל מפיק של מסמך-Word הקשור-להחלטה חייב להחיל עיצוב דרך הסגנונות שמוגדרים
+בטמפלט היחיד (skills/docx/decision_template.docx), ולא לקבוע פונט/גודל ידנית
+(``rFonts`` / ``sz`` / ``szCs``) — אלה נגזרים מ-styles.xml של הטמפלט.
+
+הטסט גורף את מודולי-המפיקים ונכשל אם נבנה אלמנט ``w:rFonts`` / ``w:sz`` /
+``w:szCs`` בקוד, אלא אם השורה נושאת את הסמן המפורש ``# INV-EX4-ok`` (חריג
+מתועד — למשל חיזוק גופן-התבנית עצמו David כעקיפת באג-RTL של Word).
+
+ספ: docs/spec/06-export.md §INV-EX4.
+"""
+
+from __future__ import annotations
+
+import re
+from pathlib import Path
+
+import pytest
+
+from legal_mcp.services import (
+    analysis_docx_exporter,
+    docx_exporter,
+    docx_reviser,
+    docx_retrofit,
+)
+
+# Every module that builds/edits a decision-related Word document.
+_PRODUCER_MODULES = [
+    docx_exporter,
+    analysis_docx_exporter,
+    docx_reviser,
+    docx_retrofit,
+]
+
+# Construction (not lookup) of a hand-set font/size element:
+#   OxmlElement("w:rFonts")          (python-docx producers)
+#   etree.SubElement(rPr, _w("sz"))  (raw-lxml producer: docx_reviser)
+_HANDSET_RE = re.compile(
+    r'OxmlElement\(\s*["\']w:(?:rFonts|sz|szCs)["\']'
+    r'|SubElement\([^)]*_w\(\s*["\'](?:rFonts|sz|szCs)["\']'
+)
+
+# Producers that build a document FROM SCRATCH must load the single template.
+# (docx_reviser/docx_retrofit operate on an already-templated existing doc.)
+_FROM_SCRATCH_MODULES = [docx_exporter, analysis_docx_exporter]
+
+
+def _module_path(module) -> Path:
+    return Path(module.__file__)
+
+
+def test_no_handset_font_or_size_in_producers() -> None:
+    """INV-EX4: no hand-set rFonts/sz/szCs in any producer (only via the
+    template's named styles), unless explicitly marked ``# INV-EX4-ok``."""
+    violations: list[str] = []
+    for module in _PRODUCER_MODULES:
+        path = _module_path(module)
+        for lineno, line in enumerate(path.read_text(encoding="utf-8").splitlines(), 1):
+            if _HANDSET_RE.search(line) and "INV-EX4-ok" not in line:
+                violations.append(f"{path.name}:{lineno}: {line.strip()}")
+    assert not violations, (
+        "INV-EX4 violation — hand-set font/size outside the template styles "
+        "(use a named template style, or mark with # INV-EX4-ok):\n  "
+        + "\n  ".join(violations)
+    )
+
+
+@pytest.mark.parametrize("module", _FROM_SCRATCH_MODULES)
+def test_from_scratch_producers_load_the_single_template(module) -> None:
+    """INV-EX4: from-scratch producers reference the one template file."""
+    source = _module_path(module).read_text(encoding="utf-8")
+    assert "decision_template.docx" in source, (
+        f"{_module_path(module).name} must load the single template "
+        "skills/docx/decision_template.docx (INV-EX4)."
+    )

mcp-server/tests/test_protocol_analyzer.py

@@ -49,6 +49,65 @@ def test_find_protocol_none():
     assert pa._find_protocol(docs) is None
 
 
+def test_find_protocol_skips_lower_scope():
+    # A local/district-committee protocol (scope='lower') must NOT be picked for
+    # the hearing-vs-pleadings comparison; the ועדת-ערר one wins (#223).
+    docs = [
+        {"id": "lower", "doc_type": "protocol", "title": "נספח 18 — פרוטוקולי ועדה מקומית",
+         "metadata": {"protocol_scope": "lower"}, "created_at": "2026-01-01"},
+        {"id": "appeal", "doc_type": "protocol", "title": "פרוטוקול דיון 23.6.26",
+         "metadata": {}, "created_at": "2026-06-23"},
+    ]
+    assert pa._find_protocol(docs)["id"] == "appeal"
+
+
+def test_find_protocol_all_lower_returns_none():
+    # If every protocol is lower-scoped there is no ערר-hearing to compare.
+    docs = [
+        {"id": "1", "doc_type": "protocol", "title": "פרוטוקול מקומי",
+         "metadata": {"protocol_scope": "lower"}},
+    ]
+    assert pa._find_protocol(docs) is None
+
+
+def test_find_protocol_prefers_most_recent_appeal():
+    docs = [
+        {"id": "old", "doc_type": "protocol", "title": "פרוטוקול א",
+         "metadata": {}, "created_at": "2026-03-01"},
+        {"id": "new", "doc_type": "protocol", "title": "פרוטוקול ב",
+         "metadata": {}, "created_at": "2026-06-23"},
+    ]
+    assert pa._find_protocol(docs)["id"] == "new"
+
+
+def test_find_protocol_explicit_document_id_overrides():
+    # Explicit target wins even over scope/recency heuristics — including the
+    # ability to point at a lower-scoped doc if the caller insists.
+    docs = [
+        {"id": "aaaaaaaa-0000-0000-0000-000000000001", "doc_type": "protocol",
+         "title": "פרוטוקול ערר", "metadata": {}, "created_at": "2026-06-23"},
+        {"id": "aaaaaaaa-0000-0000-0000-000000000002", "doc_type": "protocol",
+         "title": "פרוטוקול מקומי", "metadata": {"protocol_scope": "lower"},
+         "created_at": "2026-01-01"},
+    ]
+    from uuid import UUID
+    picked = pa._find_protocol(
+        docs, document_id=UUID("aaaaaaaa-0000-0000-0000-000000000002"),
+    )
+    assert picked["id"] == "aaaaaaaa-0000-0000-0000-000000000002"
+
+
+def test_find_protocol_unknown_document_id_returns_none():
+    from uuid import UUID
+    docs = [
+        {"id": "aaaaaaaa-0000-0000-0000-000000000001", "doc_type": "protocol",
+         "title": "פרוטוקול", "metadata": {}},
+    ]
+    assert pa._find_protocol(
+        docs, document_id=UUID("bbbbbbbb-0000-0000-0000-000000000009"),
+    ) is None
+
+
 # ── _normalize_change — anti-hallucination gate (INV-AH) ───────────────────
 
 _AID = "11111111-1111-1111-1111-111111111111"

mcp-server/tests/test_track_changes_e2e.py

@@ -207,8 +207,9 @@ def test_multiple_revisions_all_tracked_independently(tmp_path: Path) -> None:
 
 
 def test_rtl_preserved_in_tracked_insertion(tmp_path: Path) -> None:
-    """Inserted paragraph must have bidi + rtl + David font so it renders
-    correctly in Word alongside the user's content."""
+    """Inserted paragraph must have bidi + an <w:rtl/> run marker so Hebrew
+    renders via the template's cs=David slot — but font/size come from the
+    template's named paragraph style (INV-EX4), not hand-set rFonts/sz."""
     user_file = tmp_path / "user.docx"
     _make_user_edited_docx(user_file)
     docx_retrofit.retrofit_bookmarks(user_file, backup=False)
@@ -229,9 +230,10 @@ def test_rtl_preserved_in_tracked_insertion(tmp_path: Path) -> None:
             if any("עברית RTL" in (t.text or "") for t in text_els):
                 rPr = r.find(_w("rPr"))
                 assert rPr is not None
+                # RTL marker preserved…
                 assert rPr.find(_w("rtl")) is not None
-                rFonts = rPr.find(_w("rFonts"))
-                assert rFonts is not None
-                assert rFonts.get(_w("ascii")) == "David"
+                # …but NO hand-set font/size on the run (INV-EX4).
+                assert rPr.find(_w("rFonts")) is None
+                assert rPr.find(_w("sz")) is None
                 return
     pytest.fail("tracked insertion with 'עברית RTL' not found")

scripts/SCRIPTS.md

@@ -76,7 +76,8 @@
 | `calibrate_halacha_dedup.py` | python | **#82.1** — כיול ספי ה-dedup הלקסיקלי (#82.3) מול gold-set הניקוי. קורא `halacha-cleanup-manifest-*.csv` (זוגות duplicate↔survivor מתויגי-אדם), טוען טקסט-survivor מה-DB, ו-sweep של (jaccard_min × levenshtein_min) עם P/R/F1, מסמן את נקודת-העבודה המוגדרת. אימת ש-(0.55, 0.70) → **precision 1.0** (אפס false-merge), recall 0.30 — מתאים לאיתות-משני שחוסם auto-approve. `--manifest <path>`. רץ עם venv של mcp-server | חד-פעמי — כיול (בוצע 2026-06-06) |
 | `ab_halacha_opus48.py` | python | **A/B לא-הרסני לחילוץ הלכות (Claude)** — מריץ מחדש חילוץ הלכות על פסק-דין בודד דרך מודל/effort נבחרים (`AB_MODEL`/`AB_EFFORT`, ברירת-מחדל `claude-opus-4-8`/`xhigh`) ומשווה לסטטיסטיקות ההלכות הקיימות ב-DB **בלי למחוק/לכתוב כלום**. משכפל את `halacha_extractor.extract()` (אותם פרומפטים, בחירת-צ'אנקים, אימות-ציטוט) ומחליף רק את קריאת ה-LLM ב-`claude -p --model --effort`. מפיק `data/ab_halacha_<case>_<effort>.json`. הרצה: `DOTENV_PATH=/home/chaim/.env DATA_DIR=.../data .venv/bin/python scripts/ab_halacha_opus48.py <case_law_id>`. **ממצא 2026-05-31 (שטיין 1128-08-20):** Opus 4.8@xhigh חילץ 51 מול 124 בייצור (100% quote-verified מול 96%) אך ביטחון מכויל-נמוך יותר (חציון 0.75 מול 0.82) — ולכן **לא** מקטין את תור-האישור-הידני תחת sweep אוטו-אישור conf≥0.78 (26 מול 24). שיפור איכות, לא צמצום-תור. | ידני (החלטת מודל-חילוץ) |
 | `ab_halacha_codex.py` | python | **A/B לא-הרסני לחילוץ הלכות (Codex/gpt-5.5)** — עמית ל-`ab_halacha_opus48` אך מחליף את `claude -p` ב-`codex exec --model gpt-5.5` (אימות ChatGPT, ללא OPENAI_API_KEY). אותם פרומפטים ואותו הסקת quote-verification. הפלט האחרון של הסוכן (`-o FILE`) נפענח כ-JSON. `AB_MODEL` (default `gpt-5.5`), `AB_REASONING` low/medium/high/xhigh (default `medium`), `AB_CONCURRENCY` (default 1), `CODEX_BIN`. מפיק `data/ab_halacha_codex_<case>_<model>_<reasoning>.json`. הרצה: `DOTENV_PATH=/home/chaim/.env DATA_DIR=.../data mcp-server/.venv/bin/python scripts/ab_halacha_codex.py <case_law_id>`. **ממצא 2026-06-17 (8181-21 האוניברסיטה העברית):** gpt-5.5@medium חילץ 27 מול 28 של Opus (quote-verified 100%/100%), ביטחון חציון 0.86 מול 0.78 — אך **0 פריטים מתחת ל-0.7** (לעומת 9/28 של Opus = 32%), דבר המצביע על over-confidence. holding↑ (12 מול 7), procedural↓ (4 מול 7). **מסקנה: ריאלי כ-fallback חירום; לא מוכן לייצור ללא כיול-ביטחון.** | ידני (בנצ'מרק מודל codex) |
-| `calibrate_effort.py` | python | **#208 (WS5/Q1, INV-G8 eval-harness) — כיול model×effort של הכותב מול הסופיים.** A/B per-(תיק,בלוק,effort) על `draft_final_pairs` בעלי `final_text`: מייצר מחדש כל בלוק דרך מסלול-הייצור (`block_writer.write_block(effort_override=…)` → `claude_session.query` → `claude -p`, Opus 4.8 נעוץ, **מקומי-בלבד**) ומודד מול ה**סקשן** המתאים בסופי דרך `services/style_distance.block_distance_to_final` (מקור-מדידה יחיד, G2): `change_percent` (compute_diff_stats) · `anti_pattern_total` (`lessons.ANTI_PATTERNS`, הסיגנל הנקי-לסגנון) · `golden_ratio_deviation_pp` · `distance` מרוכב. **ממליץ** per-בלוק על ה-effort בעל ה-distance-הממוצע-הנמוך (לצד ברירת-המחדל מ-#204). מפיק `data/eval/effort-calibration-<ts>.{json,md}`. ⚠️ **גודל-מדגם מודפס בראש הדוח — עדות-כיוון, לא רגרסיה** (מעט סופיים-עלויים). `--self-test` (offline, אפס DB/CLI — מוכיח את לוגיקת-המדידה) · `--dry-run` (תכנון-גריד) · `--efforts`/`--blocks`/`--case`/`--repeats`. **חובה מקומי** (claude CLI; לא בקונטיינר/worktree-ללא-CLI). הרצה: `POSTGRES_PASSWORD=… mcp-server/.venv/bin/python scripts/calibrate_effort.py`. | ידני — לכיול ברירות-effort של הכותב |
+| `calibrate_effort.py` | python | **#208 (WS5/Q1, INV-G8 eval-harness) — כיול model×effort של הכותב מול הסופיים.** A/B per-(תיק,בלוק,effort) על `draft_final_pairs` בעלי `final_text`: מייצר מחדש כל בלוק דרך מסלול-הייצור (`block_writer.write_block(effort_override=…)` → `claude_session.query` → `claude -p`, Opus 4.8 נעוץ, **מקומי-בלבד**) ומודד מול ה**סקשן** המתאים בסופי דרך `services/style_distance.block_distance_to_final` (מקור-מדידה יחיד, G2): `change_percent` (compute_diff_stats) · `anti_pattern_total` (`lessons.ANTI_PATTERNS`, הסיגנל הנקי-לסגנון) · `golden_ratio_deviation_pp` · `distance` מרוכב. **ממליץ** per-בלוק (style-clean, #213) לפי `anti_pattern_total` **ראשי** (סיגנל-הסגנון הנקי) → סטיית-יחסי-זהב → `distance` (tiebreak קלוש); **`change_percent` מדווח-לא-מדורג** (מערבב סגנון עם שלמות-תוכן, 07-learning §0.7). ההמלצה נושאת דגל-`confidence` ("clear"/"weak" — הובלת-anti < epsilon=max(0.5, 0.20·spread) ⇒ ⚠️weak בתוך-הרעש, לא מוצג כוודאי) לצד ברירת-המחדל מ-#204. מפיק `data/eval/effort-calibration-<ts>.{json,md}` ב**כתיבה אינקרמנטלית** — הדוח נכתב מחדש אחרי כל בלוק שמסתיים, כך שקריסה באמצע הגריד לא מאבדת בלוקים שכבר יוצרו (INV-G8). כל תא-(תיק,בלוק,effort) שנכשל (rate-limit חולף, או חריגה) נרשם כ-warning ומדולג — לא מפיל את הריצה. **`block-yod` (דיון) הוצא מברירות-המחדל** (`CALIBRATABLE_BLOCKS`) — דורש כיוון-מאושר (brainstorm→approve_direction) שתיקי-הכיול חסרים ולכן `write_block` זורק; נשאר ב-`VALID_BLOCKS` לכפייה ידנית (`--blocks block-yod`) על-אחריות-המשתמש, כשהשער-per-תא שומר אותו לא-פטאלי. ⚠️ **גודל-מדגם מודפס בראש הדוח — עדות-כיוון, לא רגרסיה** (מעט סופיים-עלויים). `--self-test` (offline, אפס DB/CLI — מוכיח מדידה + דירוג-סגנון + confidence + rerank) · **`--rerank <report.json>`** (offline, אפס LLM/DB — מדרג-מחדש דוח-כיול שמור תחת המפתח החדש, מדפיס טבלה + כותב `*-reranked.md`) · `--dry-run` (תכנון-גריד) · `--efforts`/`--blocks`/`--case`/`--repeats`. **חובה מקומי** (claude CLI; לא בקונטיינר/worktree-ללא-CLI). הרצה: `POSTGRES_PASSWORD=… mcp-server/.venv/bin/python scripts/calibrate_effort.py`. | ידני — לכיול ברירות-effort של הכותב |
+| `calibrate_block_yod.py` | python | **#212 (WS5/Q1, INV-G8 eval-harness) — כיול effort ל-בלוק י' (דיון והכרעה) בלבד.** ה-harness ש-#208 לא יכל לכייל: `write_block(block-yod)` זורק "ללא כיוון מאושר" כי בלוק-הדיון דורש `decisions.direction_doc.approved` שתיקי-הכיול חסרים. **התוספת על #208 — שכבת אספקת-כיוון read-only:** לפני ייצור בלוק-הדיון, ה-harness מזרים `direction_doc` שמיש דרך **overlay על קריאת-ה-DB** (monkeypatch של `db.get_decision_by_case`/`db.get_decision` לתהליך-הזה-בלבד) — **אפס כתיבה ל-DB**, מצב-תיק-הייצור לא משתנה (השער reverted ב-finally). **אסטרטגיית-אספקה per-תיק:** (1) **stored** (מועדף, נאמן) — כיוון-מאושר שמור (`approved`+`selected_direction`, או סכמת-8126 עם `direction_id`/`structure`); (2) **reconstructed** (נסיגה, עדות-כיוון-בלבד) — `direction_doc` מינימלי שנגזר מ-`decision.outcome` + סקשן-הדיון של הסופי **באותו תיק** (INV-LRN5: outcome+מצביע-לסופי-עצמי, אפס מהות חוצת-תיקים, מתויג `reconstructed`); (3) **skip** אם אין דיון/כיוון. מודד דרך אותו `style_distance.block_distance_to_final` (מקור-מדידה יחיד, G2) ומשתמש-חוזר ב-`_score_cell`/`recommend_effort` מ-`calibrate_effort` (#208). **ממצא-קורפוס (2026-06-30):** מ-8 סופיים, 7 ניתנים-לכיול; רק 8126-03-25 הוא stored+scorable (8174-12-24 שמור-אך-חסר-סקשן-דיון) → `--allow-reconstructed false` נותן תא-יחיד; ברירת-המחדל (reconstructed פעיל) נותנת 7 → הנסיגה היא מה שמפעיל את ה-harness היום. כתיבה אינקרמנטלית per-תיק + שער-per-תא לא-פטאלי (INV-G8). ⚠️ **גודל-מדגם בראש הדוח — עדות-כיוון, לא רגרסיה.** `--self-test` (offline, אפס DB/CLI — מוכיח מדידה + לוגיקת-אספקת-כיוון + overlay) · `--dry-run` (תכנון-גריד + מצב-אספקה per-תיק, צריך DB) · `--efforts`/`--case`/`--repeats`/`--allow-reconstructed`. ברירת-מחדל נוכחית: `xhigh` (`block_writer.py:72`, 'by feel'). **חובה מקומי** (claude CLI; לא בקונטיינר/worktree-ללא-CLI). הרצת-כיול-חי (לבצע ע"י היו"ר): `POSTGRES_PASSWORD=… POSTGRES_HOST=127.0.0.1 POSTGRES_PORT=5433 mcp-server/.venv/bin/python scripts/calibrate_block_yod.py`. | ידני — לכיול ה-effort של בלוק-הדיון (היו"ר מריץ; long, צורך-מכסה) |
 | `monitor_halacha_quality.py` | python | מנטר איכות חילוץ הלכות. בודק drift של `avg(confidence)` בין baseline היסטורי לחלון אחרון. מחזיר JSON מטריקות + alert ב-stderr אם drift > threshold (ברירת מחדל 5%). 2 סדרות: trusted (approved+published) ו-all_extracted. תומך `--window N` / `--threshold X` / `--min-sample N` / `--silent` / `--exit-on-alert`. רץ ב-container או מקומית עם `mcp-server/.venv` (אין תלות ב-LLM, רק SQL). **תזמון מומלץ**: `0 8 * * 1` (יום ראשון 08:00, שבועי) | `0 8 * * 1` (לתזמן) |
 | `audit_training_corpus.py` | python | audit של `style_corpus` — לכל החלטה: שדות מטא-דאטה מאוכלסים (`summary`/`outcome`/`key_principles`/`appeal_subtype`/`subject_categories`), קישור ל-`documents` (FK + chunks + embeddings). מפיק `data/audit/corpus-YYYY-MM-DD.json` + summary בקונסול. דרוש `POSTGRES_URL` או POSTGRES_*. אין תלויות חיצוניות מלבד asyncpg. **רץ מהמכונה המקומית** (לא קונטיינר) — חיבור ישיר ל-Postgres :5433 | ידני / קדם-עבודה לפני enrichment של מטא-דאטה |
 | `backfill_style_exemplars.py` | python | **T1 (style-acquisition)** — מאכלס `style_exemplars` מקורפוס דפנה (`style_corpus` + `internal_committee` chair=דפנה): מפצל לסעיפים (`chunker._split_into_sections`) → פסקאות (25-450 מילים) → embed (Voyage) → שמירה עם `section`/`outcome`/`practice_area`. מאפשר לכותב לאחזר פסקאות-בלוק אמיתיות של דפנה (T2/T3). מקור-סגנון בלבד (INV-LRN5). אידמפוטנטי (מנקה per-decision). `--dry-run` (default) / `--apply`. דורש POSTGRES_URL + Voyage. **רץ מקומית** (venv). | ידני (`python scripts/backfill_style_exemplars.py --apply`) |
@@ -174,6 +175,7 @@
 
 | Script | Original Purpose | Superseded By |
 |--------|-----------------|---------------|
+| `exec_summary_1043.py` | חד-פעמי — רינדור DOCX של סיכום-מנהלים לתיק 1043-02-26 דרך עוזרי-התבנית של `analysis_docx_exporter` (פרוטוטיפ). | `analysis_docx_exporter.build_party_claims_summary_docx()` + endpoint `GET /api/cases/{n}/research/party-claims-summary/export-docx` (WS3 #202; תבניתי, INV-EX4) |
 | `backfill_pattern_frequency.py` | עדכון תדירות דפוסי סגנון ב-DB | `web/app.py::_extract_pattern_variants()` |
 | `batch_upload_training.py` | העלאת קורפוס אימון (16 קבצים) | Web UI: `/api/training/upload` |
 | `benchmark_embeddings.py` | השוואת מודלי embeddings (voyage-3 vs voyage-4) | הושלם — voyage-3-large נבחר |

scripts/backfill_case_status_trim.py

@@ -12,8 +12,6 @@ Mapping (removed → kept):
 
     uploading          → processing
     in_progress        → outcome_set
-    analyst_verified   → documents_ready
-    research_complete  → documents_ready
     brainstorming      → outcome_set
     analysis_enriched  → direction_approved
     ready_for_writing  → direction_approved
@@ -50,12 +48,14 @@ from legal_mcp.services import db  # noqa: E402
 logging.basicConfig(level=logging.INFO, format="%(asctime)s [%(levelname)s] %(message)s")
 log = logging.getLogger("status-trim")
 
-# removed status → nearest preceding kept status
+# removed status → nearest preceding kept status.
+# NOTE (2026-06-30): analyst_verified + research_complete were RE-CANONICALISED
+# into the status model (legal_mcp/case_status_model.py, "thinking" phase) — they
+# are valid statuses again, so they were removed from this map. Do NOT downgrade
+# them; only genuinely-removed markers remain below.
 STATUS_MAP = {
     "uploading":         "processing",
     "in_progress":       "outcome_set",
-    "analyst_verified":  "documents_ready",
-    "research_complete": "documents_ready",
     "brainstorming":     "outcome_set",
     "analysis_enriched": "direction_approved",
     "ready_for_writing": "direction_approved",

scripts/calibrate_block_yod.py

@@ -0,0 +1,646 @@
+#!/usr/bin/env python3
+"""#212 (WS5 / Q1, INV-G8 eval-harness) — DEDICATED effort calibration for בלוק י'
+(block-yod, the discussion/decision block — דיון והכרעה).
+
+WHY A SEPARATE HARNESS (the #208 harness could NOT calibrate block-yod):
+  block-yod is the most important block — its effort is currently `xhigh`, set
+  "by feel" (block_writer.py:72), never empirically calibrated. The general #208
+  harness (scripts/calibrate_effort.py) EXCLUDES it from CALIBRATABLE_BLOCKS
+  because block_writer.write_block(case_id, "block-yod", …) raises
+      ValueError("לא ניתן לכתוב בלוק דיון ללא כיוון מאושר…")
+  — the discussion block requires an APPROVED DIRECTION (brainstorm →
+  approve_direction → decisions.direction_doc{"approved": True, …}) that the
+  calibration cases (the draft_final_pairs finals) do not have. So block-yod is
+  not calibratable standalone without first SUPPLYING a direction.
+
+WHAT THIS HARNESS ADDS over #208 — the DIRECTION-SUPPLY layer (read-only):
+  Before regenerating block-yod for a calibration case, it makes a usable
+  `direction_doc` available to write_block via a READ-ONLY DB OVERLAY (it
+  monkeypatches db.get_decision_by_case / db.get_decision FOR THIS PROCESS ONLY
+  to overlay the direction into the returned decision dict). NOTHING is written
+  to the DB — production case state is never mutated (no UPDATE), satisfying the
+  "do not pollute real case state" constraint. The overlay is reverted in a
+  finally block.
+
+  DIRECTION-SUPPLY STRATEGY (per-case, documented in the report):
+    1. STORED (preferred, most faithful): if the case's decision already carries
+       a usable approved direction (direction_doc.approved is truthy and it has a
+       selected_direction, OR the 8126-style schema with direction_id/structure),
+       use it verbatim. This is the production-faithful path.
+    2. RECONSTRUCTED (fallback, directional-only): otherwise synthesize a MINIMAL
+       direction_doc from the case's decision.outcome + the final's OWN discussion
+       section (the chair's signed reasoning is the ground truth for "what
+       direction was taken"). Clearly LABELLED reconstructed in the report → the
+       resulting numbers are directional, not faithful to a real brainstorm.
+       INV-LRN5: the reconstructed direction carries the case's OWN outcome +
+       a neutral pointer to its OWN final — no substance dragged across cases.
+
+  EMPIRICAL NOTE (measured on the current corpus, 2026-06-30): of the 8 finals,
+  only 8174-12-24 has a STORED approved direction — and its final has NO parsable
+  discussion section (0 words), so it is not scorable for block-yod anyway. The 7
+  finals that DO have a scorable discussion section all LACK a usable stored
+  direction. Hence the RECONSTRUCTED fallback is what makes this harness useful
+  today; --dry-run prints the per-case supply mode so the operator sees this.
+
+WHAT IT MEASURES (identical to #208, REUSED — no parallel metric path, G2):
+  Regenerate block-yod at each effort via block_writer.write_block(effort_override)
+  (the PRODUCTION generation path → claude_session.query → `claude -p`, pinned
+  Opus 4.8, local-only), then score vs the final's DISCUSSION section via
+  services.style_distance.block_distance_to_final (change_percent /
+  anti_pattern_total / golden_ratio_deviation_pp / composite distance). The
+  recommend_effort, aggregate_cell, _score_cell and direction-free scaffolding are
+  imported from calibrate_effort (#208) — this file only supplies the direction
+  and restricts the grid to block-yod.
+
+GENERATION PATH (do not violate — reference_claude_generation_path /
+claude_session docstring): write_block → claude_session.query → `claude -p` uses
+the local claude.ai session. It runs ONLY on the host where the `claude` CLI
+exists (NOT the legal-ai container, NOT a symlinked worktree without CLI access).
+Hence the live A/B is HOST-ONLY; --self-test proves the measurement +
+direction-supply logic offline with zero model calls.
+
+⚠️ SAMPLE-SIZE CAVEAT (honored, not hidden): very few cases have an uploaded
+final WITH a parsable discussion section. The report prints n_finals PROMINENTLY
+and labels the output DIRECTIONAL EVIDENCE, not a regression. With n<3 the
+recommendation is advisory only; the chair/operator decides whether to adopt.
+
+Usage (mcp-server venv; live needs POSTGRES + the `claude` CLI on the host):
+    PY=/home/chaim/legal-ai/mcp-server/.venv/bin/python
+    $PY scripts/calibrate_block_yod.py --self-test          # offline proof, no DB/CLI
+    $PY scripts/calibrate_block_yod.py --dry-run            # plan grid + supply modes (needs DB, no model)
+    POSTGRES_PASSWORD=… POSTGRES_HOST=127.0.0.1 POSTGRES_PORT=5433 \
+        $PY scripts/calibrate_block_yod.py                  # live A/B over finals (host-only)
+    … --efforts low,medium,high,xhigh                       # override the effort grid
+    … --case 1130-08-25                                     # a single case
+    … --repeats 2                                           # avg N gens/cell (noise)
+    … --allow-reconstructed false                           # STORED-direction cases only (faithful)
+"""
+from __future__ import annotations
+
+import argparse
+import asyncio
+import json
+import logging
+import os
+import sys
+from datetime import datetime
+from pathlib import Path
+from statistics import mean
+
+logger = logging.getLogger(__name__)
+
+REPO_ROOT = Path(__file__).resolve().parent.parent
+sys.path.insert(0, str(REPO_ROOT / "mcp-server" / "src"))
+sys.path.insert(0, str(REPO_ROOT / "scripts"))
+
+if "POSTGRES_URL" not in os.environ:
+    os.environ["POSTGRES_URL"] = (
+        f"postgres://{os.environ.get('POSTGRES_USER','legal_ai')}:"
+        f"{os.environ.get('POSTGRES_PASSWORD','')}@"
+        f"{os.environ.get('POSTGRES_HOST','127.0.0.1')}:"
+        f"{os.environ.get('POSTGRES_PORT','5433')}/"
+        f"{os.environ.get('POSTGRES_DB','legal_ai')}"
+    )
+
+# Reuse the #208 machinery wholesale (G2 — no parallel harness/metric path).
+# These are pure/offline-safe imports; calibrate_effort does no work at import.
+from calibrate_effort import (  # noqa: E402
+    IL_TZ,
+    OUT_DIR,
+    VALID_EFFORTS,
+    _current_default,
+    _finals_for_calibration,
+    _score_cell,
+    recommend_effort,
+)
+
+BLOCK_ID = "block-yod"
+SECTION = "discussion"  # block-yod's golden-ratio section (style_distance._BLOCK_TO_SECTION)
+DEFAULT_EFFORTS = ["low", "medium", "high", "xhigh"]
+
+
+# ── direction supply (the block-yod-specific piece) ──────────────────────────
+def _stored_direction_is_usable(direction_doc: dict | None) -> bool:
+    """True if the case's stored direction_doc can drive write_block(block-yod).
+
+    write_block gates on `direction_doc.get("approved")` being truthy. Two stored
+    schemas exist in the corpus:
+      • brainstorm schema (build_direction_doc): {"approved": True, "selected_direction": …}
+      • chair-approved schema (8126-style): {"direction_id", "direction_name",
+        "structure", "template", …} — has NO "approved" key, so write_block would
+        reject it as-is. We treat it as usable IFF it carries a concrete direction
+        (direction_id/direction_name), and _normalize_stored_direction() stamps
+        approved=True onto the OVERLAY copy (never the DB).
+    Empty / brainstorm-only ({"brainstorm": …, "approved": False}) → not usable.
+    """
+    dd = direction_doc or {}
+    if dd.get("approved") and (dd.get("selected_direction") or dd.get("reasoning")):
+        return True
+    if dd.get("direction_id") or dd.get("direction_name"):
+        return True
+    return False
+
+
+def _normalize_stored_direction(direction_doc: dict) -> dict:
+    """Return an overlay copy of a usable stored direction with approved=True.
+
+    Pure dict transform on a COPY — the real decisions.direction_doc is untouched.
+    The 8126-style schema lacks `approved`; stamping it here lets write_block's gate
+    pass without us mutating the DB.
+    """
+    dd = dict(direction_doc)
+    dd["approved"] = True
+    return dd
+
+
+def reconstruct_minimal_direction(outcome: str, final_discussion: str) -> dict:
+    """Build a MINIMAL, clearly-reconstructed direction_doc (fallback path).
+
+    Faithful enough to make write_block(block-yod) runnable, honest about being a
+    reconstruction: the chair's SIGNED final discussion is the ground truth for
+    "what direction was taken", and the case's OWN outcome fixes the structure.
+    INV-LRN5: case's own outcome + a neutral pointer to its own final only — no
+    cross-case substance. Pure (no DB/LLM) → unit-tested in --self-test.
+    """
+    from legal_mcp.services.lessons import OUTCOME_LABELS_HE, canonical_outcome
+
+    oc = canonical_outcome(outcome or "rejection")
+    snippet = " ".join((final_discussion or "").split()[:120])
+    return {
+        "approved": True,
+        "reconstructed": True,  # marker so the report can label it
+        "outcome": oc,
+        "outcome_hebrew": OUTCOME_LABELS_HE.get(oc, oc),
+        "reasoning": (
+            "כיוון משוחזר לצורך כיול-effort בלבד (לא סיעור-מוחות אמיתי): "
+            "התוצאה והכיוון נגזרים מההחלטה החתומה של היו\"ר בתיק זה. "
+            "תמצית פתיח-הדיון של הסופי: " + snippet
+        ),
+        "selected_direction": {
+            "name": f"כיוון משוחזר — {OUTCOME_LABELS_HE.get(oc, oc)}",
+            "reasoning": [
+                "שחזר את שלד-הנימוק מהדיון החתום של היו\"ר בתיק זה.",
+            ],
+            "precedents": [],
+        },
+        "additional_notes": (
+            "⚠️ כיוון משוחזר — תוצאת-הכיול היא עדות-כיוון, לא נאמנה למסמך-כיוון אמיתי."
+        ),
+    }
+
+
+def plan_direction_supply(
+    decision: dict | None,
+    outcome: str,
+    final_discussion: str,
+    allow_reconstructed: bool,
+) -> tuple[dict | None, str]:
+    """Decide how to supply block-yod's direction for one case.
+
+    Returns (overlay_direction_doc | None, mode) where mode ∈
+    {"stored", "reconstructed", "skip"}. Pure (no I/O) → unit-tested.
+    """
+    stored = (decision or {}).get("direction_doc")
+    if _stored_direction_is_usable(stored):
+        return _normalize_stored_direction(stored), "stored"
+    if allow_reconstructed and (final_discussion or "").strip():
+        return reconstruct_minimal_direction(outcome, final_discussion), "reconstructed"
+    return None, "skip"
+
+
+class _DirectionOverlay:
+    """Read-only overlay: make db.get_decision_by_case / db.get_decision return the
+    case's decision WITH `direction_doc` replaced by the supplied overlay — for the
+    target case_id only, for the duration of one cell. Reverted in __exit__.
+
+    write_block re-fetches the decision internally (block_writer.py:383), so passing
+    a direction dict in isn't enough — the DB read must yield it. We patch the read,
+    not the write: ZERO DB mutation (no UPDATE ever issued).
+    """
+
+    def __init__(self, case_id, direction_doc: dict):
+        self._case_id = str(case_id)
+        self._dir = direction_doc
+        self._orig_by_case = None
+        self._orig_by_id = None
+
+    def __enter__(self):
+        from legal_mcp.services import db
+
+        self._orig_by_case = db.get_decision_by_case
+        self._orig_by_id = db.get_decision
+
+        async def patched_by_case(case_id, _orig=self._orig_by_case):
+            dec = await _orig(case_id)
+            if dec and str(case_id) == self._case_id:
+                dec = dict(dec)
+                dec["direction_doc"] = self._dir
+            return dec
+
+        async def patched_by_id(decision_id, _orig=self._orig_by_id):
+            dec = await _orig(decision_id)
+            if dec and dec.get("case_id") == self._case_id:
+                dec = dict(dec)
+                dec["direction_doc"] = self._dir
+            return dec
+
+        db.get_decision_by_case = patched_by_case
+        db.get_decision = patched_by_id
+        return self
+
+    def __exit__(self, *exc):
+        from legal_mcp.services import db
+
+        if self._orig_by_case is not None:
+            db.get_decision_by_case = self._orig_by_case
+        if self._orig_by_id is not None:
+            db.get_decision = self._orig_by_id
+        return False
+
+
+async def _score_block_yod_cell(case_id, effort, final_section, final_total_words,
+                                outcome, repeats, direction_doc):
+    """Score one (case, block-yod, effort) cell with the direction overlaid.
+
+    Wraps the #208 _score_cell (REUSED) inside the read-only direction overlay so
+    write_block's approved-direction gate passes without DB mutation.
+    """
+    with _DirectionOverlay(case_id, direction_doc):
+        return await _score_cell(
+            case_id, BLOCK_ID, effort, final_section, final_total_words, outcome, repeats,
+        )
+
+
+# ── self-test (no DB, no model) ──────────────────────────────────────────────
+def _self_test() -> int:
+    ok = True
+
+    def chk(name, cond):
+        nonlocal ok
+        ok = ok and cond
+        print(f"  {name:50} {'ok' if cond else 'FAIL'}")
+
+    # measurement reuse: block_distance_to_final on the discussion section.
+    from legal_mcp.services.style_distance import block_distance_to_final, split_final_by_section
+
+    final_disc = "אנו סבורים כי דין הערר להידחות. כידוע, הלכה פסוקה היא. " * 30
+    d_same = block_distance_to_final(BLOCK_ID, final_disc, final_disc, "rejection",
+                                     section_target_total_words=len(final_disc.split()))
+    chk("identical regen ⇒ change_percent==0", d_same["change_percent"] == 0.0)
+    chk("identical regen ⇒ section==discussion", d_same["section"] == SECTION)
+
+    clean = "אין בידנו לקבל את הטענה. יחד עם זאת, מצאנו טעם. " * 20
+    dirty = "## כותרת\n- נקודה\n- נקודה\n* עוד\n### תת\n" * 10
+    d_clean = block_distance_to_final(BLOCK_ID, clean, final_disc, "rejection")
+    d_dirty = block_distance_to_final(BLOCK_ID, dirty, final_disc, "rejection")
+    chk("dirty regen ⇒ more anti-patterns", d_dirty["anti_pattern_total"] > d_clean["anti_pattern_total"])
+    chk("dirty regen ⇒ larger distance", d_dirty["distance"] > d_clean["distance"])
+
+    # ── direction-supply logic ──
+    chk("empty direction ⇒ not usable", not _stored_direction_is_usable({}))
+    chk("brainstorm-only ⇒ not usable",
+        not _stored_direction_is_usable({"brainstorm": {"directions": [1]}, "approved": False}))
+    chk("approved+selected ⇒ usable",
+        _stored_direction_is_usable({"approved": True, "selected_direction": {"name": "x"}}))
+    chk("8126-style (direction_id) ⇒ usable",
+        _stored_direction_is_usable({"direction_id": "d1", "structure": "..."}))
+
+    src = {"direction_id": "d1", "structure": "..."}
+    norm = _normalize_stored_direction(src)
+    chk("normalize stamps approved=True", norm["approved"] is True)
+    chk("normalize does not mutate source", "approved" not in src)
+
+    rec = reconstruct_minimal_direction("rejection", final_disc)
+    chk("reconstruct ⇒ approved", rec["approved"] is True)
+    chk("reconstruct ⇒ marked reconstructed", rec["reconstructed"] is True)
+    chk("reconstruct ⇒ canonical outcome", rec["outcome"] == "rejection")
+    chk("reconstruct ⇒ has selected_direction", bool(rec.get("selected_direction")))
+    chk("reconstruct ⇒ no cross-case precedents (INV-LRN5)",
+        rec["selected_direction"]["precedents"] == [])
+
+    # _build_direction_context must accept BOTH supply shapes (write_block uses it).
+    from legal_mcp.services.block_writer import _build_direction_context
+    ctx_recon = _build_direction_context({"direction_doc": rec})
+    chk("recon direction renders (not 'לא אושר')", "כיוון לא אושר" not in ctx_recon)
+    ctx_stored = _build_direction_context({"direction_doc": norm})
+    chk("stored(8126) direction renders", "כיוון לא אושר" not in ctx_stored)
+
+    # plan_direction_supply: stored wins; else reconstructed; else skip.
+    dd_stored = {"direction_doc": {"approved": True, "selected_direction": {"name": "x"}}}
+    _, m1 = plan_direction_supply(dd_stored, "rejection", final_disc, allow_reconstructed=True)
+    chk("plan: stored-direction case ⇒ stored", m1 == "stored")
+    _, m2 = plan_direction_supply({"direction_doc": {}}, "rejection", final_disc, True)
+    chk("plan: no stored + discussion ⇒ reconstructed", m2 == "reconstructed")
+    _, m3 = plan_direction_supply({"direction_doc": {}}, "rejection", final_disc, False)
+    chk("plan: no stored + reconstructed disabled ⇒ skip", m3 == "skip")
+    _, m4 = plan_direction_supply({"direction_doc": {}}, "rejection", "", True)
+    chk("plan: no stored + no discussion ⇒ skip", m4 == "skip")
+
+    # overlay monkeypatch installs + reverts cleanly (no live DB needed).
+    import legal_mcp.services.db as real_db
+    saved_bc, saved_bi = real_db.get_decision_by_case, real_db.get_decision
+
+    async def fake_by_case(cid):
+        return {"id": "d", "case_id": "C1", "direction_doc": {"approved": False}}
+
+    async def fake_by_id(did):
+        return {"id": "d", "case_id": "C1", "direction_doc": {"approved": False}}
+
+    real_db.get_decision_by_case = fake_by_case
+    real_db.get_decision = fake_by_id
+    try:
+        ov_dir = {"approved": True, "reconstructed": True}
+
+        async def _probe():
+            with _DirectionOverlay("C1", ov_dir):
+                d1 = await real_db.get_decision_by_case("C1")
+                d2 = await real_db.get_decision_by_case("OTHER")
+                d3 = await real_db.get_decision("d")
+                return d1, d2, d3
+
+        # --self-test is dispatched at the __main__ entry, OUTSIDE asyncio.run,
+        # so a plain asyncio.run here is safe (no enclosing event loop).
+        d1, d2, d3 = asyncio.run(_probe())
+        chk("overlay: target case gets overlaid direction", d1["direction_doc"]["approved"] is True)
+        chk("overlay: non-target case untouched", d2["direction_doc"]["approved"] is False)
+        chk("overlay: get_decision path overlaid", d3["direction_doc"]["reconstructed"] is True)
+        chk("overlay: reverted after context", real_db.get_decision_by_case is fake_by_case)
+    finally:
+        real_db.get_decision_by_case = saved_bc
+        real_db.get_decision = saved_bi
+
+    # reused #208 primitives still behave.
+    r = recommend_effort([
+        {"effort": "high", "distance": 0.22, "anti_pattern_total": 2, "change_percent": 25, "n": 1},
+        {"effort": "xhigh", "distance": 0.22, "anti_pattern_total": 1, "change_percent": 22, "n": 1},
+    ])
+    chk("recommend tie → fewer anti-patterns", r["effort"] == "xhigh")
+    chk("current default for block-yod == xhigh", _current_default(BLOCK_ID) == "xhigh")
+    secs = split_final_by_section(
+        "רקע עובדתי\nהמקרקעין. " * 8 + "\n\nדיון והכרעה\nאנו סבורים. " * 8)
+    chk("split returns golden sections only",
+        set(secs).issubset({"background", "claims", "discussion", "summary"}))
+
+    print("ALL PASS" if ok else "*** FAILURES ***")
+    return 0 if ok else 1
+
+
+# ── live A/B (host-only — needs DB + `claude` CLI) ───────────────────────────
+async def _gather_cases(case_filter, allow_reconstructed):
+    """Build per-case metadata + direction-supply plan over the finals."""
+    from uuid import UUID
+    from legal_mcp.services import db
+    from legal_mcp.services.lessons import canonical_outcome
+    from legal_mcp.services.style_distance import split_final_by_section
+
+    finals = await _finals_for_calibration(case_filter)
+    cases = []
+    for f in finals:
+        case = await db.get_case_by_number(f["case_number"]) if f.get("case_number") else None
+        if not case:
+            continue
+        decision = await db.get_decision_by_case(UUID(case["id"]))
+        outcome = canonical_outcome((decision or {}).get("outcome", "rejection"))
+        sections = split_final_by_section(f.get("final_text", ""))
+        final_discussion = sections.get(SECTION, "")
+        final_total_words = len((f.get("final_text", "") or "").split())
+        overlay, mode = plan_direction_supply(
+            decision, outcome, final_discussion, allow_reconstructed)
+        cases.append({
+            "case_number": f["case_number"],
+            "case_id": case["id"],
+            "outcome": outcome,
+            "final_discussion": final_discussion,
+            "final_discussion_words": len(final_discussion.split()),
+            "final_total_words": final_total_words,
+            "direction": overlay,
+            "supply_mode": mode,
+            # scorable IFF it has a discussion section AND a supplied direction.
+            "scorable": bool(final_discussion.strip() and overlay is not None),
+        })
+    return cases
+
+
+async def _run(args, ts):
+    from uuid import UUID
+
+    cases = await _gather_cases(args.case, args.allow_reconstructed)
+    scorable = [c for c in cases if c["scorable"]]
+
+    grid_summary = {
+        "block": BLOCK_ID,
+        "section": SECTION,
+        "n_finals": len(cases),
+        "n_scorable": len(scorable),
+        "current_default": _current_default(BLOCK_ID),
+        "efforts": args.efforts,
+        "repeats": args.repeats,
+        "total_generations": len(scorable) * len(args.efforts) * args.repeats,
+        "allow_reconstructed": args.allow_reconstructed,
+        "cases": [
+            {"case_number": c["case_number"], "supply_mode": c["supply_mode"],
+             "scorable": c["scorable"], "outcome": c["outcome"],
+             "discussion_words": c["final_discussion_words"]}
+            for c in cases
+        ],
+    }
+
+    if args.dry_run:
+        return {"dry_run": True, "grid": grid_summary, "by_effort": []}
+
+    per_effort_runs = {e: [] for e in args.efforts}
+    per_case = []
+    for c in scorable:
+        case_cells = []
+        for effort in args.efforts:
+            # Per-cell guard (INV-G8): any single (case, effort) failure — a
+            # transient rate-limit, a too-large prompt, an unexpected raise — is
+            # logged and skipped, never allowed to abort the whole run.
+            try:
+                cell = await _score_block_yod_cell(
+                    UUID(c["case_id"]), effort, c["final_discussion"],
+                    c["final_total_words"], c["outcome"], args.repeats, c["direction"],
+                )
+            except Exception as exc:  # noqa: BLE001 — harness must survive any cell failure
+                logger.warning(
+                    "block-yod cell skipped: case=%s effort=%s mode=%s — %s",
+                    c["case_number"], effort, c["supply_mode"], exc,
+                )
+                continue
+            per_effort_runs[effort].append(cell)
+            case_cells.append({k: cell[k] for k in
+                               ("effort", "distance", "anti_pattern_total",
+                                "change_percent", "golden_ratio_deviation_pp", "n")})
+        per_case.append({"case_number": c["case_number"],
+                         "supply_mode": c["supply_mode"], "cells": case_cells})
+
+        # Incremental persistence (INV-G8): flush after EACH case so a crash later
+        # never loses cases already generated. _write_report tolerates partial rows.
+        try:
+            _write_report({"dry_run": False, "grid": grid_summary,
+                           "by_effort": _effort_rows(per_effort_runs, args.efforts),
+                           "per_case": per_case,
+                           "recommended": (recommend_effort(_effort_rows(per_effort_runs, args.efforts)) or {}).get("effort")},
+                          ts)
+        except Exception as exc:  # noqa: BLE001 — a write hiccup must not abort the run
+            logger.warning("incremental report write failed after case=%s — %s",
+                           c["case_number"], exc)
+
+    effort_rows = _effort_rows(per_effort_runs, args.efforts)
+    rec = recommend_effort(effort_rows)
+    return {"dry_run": False, "grid": grid_summary, "by_effort": effort_rows,
+            "per_case": per_case, "recommended": rec["effort"] if rec else None}
+
+
+def _effort_rows(per_effort_runs, efforts):
+    """Mean each metric across cases for each effort → one comparable row/effort."""
+    rows = []
+    for effort in efforts:
+        cells = per_effort_runs[effort]
+        if not cells:
+            continue
+        ratios = [c["golden_ratio_deviation_pp"] for c in cells
+                  if c.get("golden_ratio_deviation_pp") is not None]
+        rows.append({
+            "effort": effort,
+            "distance": round(mean(c["distance"] for c in cells), 4),
+            "anti_pattern_total": round(mean(c["anti_pattern_total"] for c in cells), 2),
+            "change_percent": round(mean(c["change_percent"] for c in cells), 2),
+            "golden_ratio_deviation_pp": round(mean(ratios), 2) if ratios else None,
+            "n": len(cells),
+        })
+    return rows
+
+
+# ── report ───────────────────────────────────────────────────────────────────
+def _ts():
+    return datetime.now(IL_TZ).strftime("%Y%m%dT%H%M%S-IL")
+
+
+def _write_report(result, ts):
+    OUT_DIR.mkdir(parents=True, exist_ok=True)
+    jp = OUT_DIR / f"block-yod-calibration-{ts}.json"
+    mp = OUT_DIR / f"block-yod-calibration-{ts}.md"
+    jp.write_text(json.dumps(result, ensure_ascii=False, indent=2), encoding="utf-8")
+
+    g = result["grid"]
+    il_now = datetime.now(IL_TZ).strftime("%Y-%m-%d %H:%M")
+    n_recon = sum(1 for c in g["cases"] if c["supply_mode"] == "reconstructed")
+    n_stored = sum(1 for c in g["cases"] if c["supply_mode"] == "stored")
+    n_skip = sum(1 for c in g["cases"] if c["supply_mode"] == "skip")
+    lines = [
+        f"# #212 — כיול effort לבלוק י' (דיון והכרעה) מול הסופיים — {ts}\n",
+        f"> נוצר: {il_now} (שעון ישראל · Asia/Jerusalem)\n",
+        f"> ⚠️ **גודל-מדגם: {g['n_scorable']} תיקים ניתנים-לכיול** "
+        f"(מתוך {g['n_finals']} סופיים). זוהי **עדות-כיוון, לא רגרסיה**. "
+        "ההמלצה אדוויזורית; ההכרעה בידי היו\"ר/המפעיל.\n",
+        f"> אספקת-כיוון: {n_stored} מאוחסן · {n_recon} משוחזר · {n_skip} מדולג. "
+        "כיוון **משוחזר** = נגזר מתוצאת-ההחלטה + פתיח-הדיון של הסופי באותו תיק "
+        "(לא סיעור-מוחות אמיתי) → התוצאה נאמנה-לכיוון פחות מתיק עם כיוון-מאושר שמור.\n",
+        f"- בלוק: {g['block']} (section={g['section']}) · ברירת-מחדל נוכחית: "
+        f"**{g['current_default'] or '—'}** (נקבע 'by feel', block_writer.py:72)",
+        f"- efforts: {', '.join(g['efforts'])} · repeats/cell: {g['repeats']}",
+        f"- סך ייצורי-מודל: {g['total_generations']}",
+        "",
+        "## אספקת-כיוון per-תיק\n",
+        "| case | supply_mode | discussion_words | outcome | scorable |",
+        "|---|---|---|---|---|",
+    ]
+    for c in g["cases"]:
+        lines.append(
+            f"| {c['case_number']} | {c['supply_mode']} | {c['discussion_words']} | "
+            f"{c['outcome']} | {'✓' if c['scorable'] else '—'} |")
+    lines.append("")
+
+    if result.get("dry_run"):
+        lines += ["## DRY-RUN — תכנון הגריד בלבד (ללא ייצור)\n",
+                  f"- תיקים ניתנים-לכיול: **{g['n_scorable']}** · "
+                  f"ייצורים מתוכננים: **{g['total_generations']}**",
+                  ""]
+        mp.write_text("\n".join(lines) + "\n", encoding="utf-8")
+        return jp, mp
+
+    rec = result.get("recommended") or "—"
+    mark = "" if rec == g["current_default"] else " ⬅︎ שינוי מומלץ"
+    lines += [
+        f"## המלצה: effort = **{rec}**{mark}  (current = {g['current_default'] or '—'})\n",
+        "## טבלת effort (distance נמוך = קרוב יותר לדיון של דפנה)\n",
+        "| effort | distance | anti_total | change% | ratioΔpp | n |",
+        "|---|---|---|---|---|---|",
+    ]
+    for r in result.get("by_effort", []):
+        star = " ⭐" if r["effort"] == result.get("recommended") else ""
+        ratio = r["golden_ratio_deviation_pp"]
+        lines.append(
+            f"| {r['effort']}{star} | {r['distance']:.4f} | {r['anti_pattern_total']} | "
+            f"{r['change_percent']} | {ratio if ratio is not None else '—'} | {r['n']} |")
+    lines.append("")
+    lines += ["## פירוט per-תיק\n", "| case | mode | effort | distance | anti | change% |",
+              "|---|---|---|---|---|---|"]
+    for pc in result.get("per_case", []):
+        for cell in pc["cells"]:
+            lines.append(
+                f"| {pc['case_number']} | {pc['supply_mode']} | {cell['effort']} | "
+                f"{cell['distance']:.4f} | {cell['anti_pattern_total']} | {cell['change_percent']} |")
+    lines.append("")
+    lines.append("> change% מערבב סגנון עם שלמות-תוכן (07-learning §0.7); "
+                 "anti_total הוא הסיגנל הנקי-יותר לסגנון. תיקים עם supply_mode=reconstructed "
+                 "תורמים עדות-כיוון בלבד.\n")
+    mp.write_text("\n".join(lines) + "\n", encoding="utf-8")
+    return jp, mp
+
+
+async def main() -> int:
+    ap = argparse.ArgumentParser(description="#212 block-yod (discussion) effort calibration harness")
+    ap.add_argument("--self-test", action="store_true",
+                    help="offline measurement + direction-supply proof (no DB/CLI)")
+    ap.add_argument("--dry-run", action="store_true",
+                    help="plan the grid + per-case direction-supply mode, no model calls (needs DB)")
+    ap.add_argument("--efforts", default=",".join(DEFAULT_EFFORTS),
+                    help=f"comma effort grid (default {','.join(DEFAULT_EFFORTS)})")
+    ap.add_argument("--case", default=None, help="restrict to a single case_number")
+    ap.add_argument("--repeats", type=int, default=1, help="generations per cell (avg out gen noise)")
+    ap.add_argument("--allow-reconstructed", default="true",
+                    help="true|false — allow the reconstructed-direction fallback (default true). "
+                         "false = STORED-direction cases only (most faithful, may yield 0 cells).")
+    args = ap.parse_args()
+
+    logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")
+
+    args.efforts = [e.strip() for e in args.efforts.split(",") if e.strip()]
+    bad = [e for e in args.efforts if e not in VALID_EFFORTS]
+    if bad:
+        print(f"invalid effort(s): {bad}. valid: {sorted(VALID_EFFORTS)}", file=sys.stderr)
+        return 2
+    args.allow_reconstructed = str(args.allow_reconstructed).strip().lower() in ("1", "true", "yes", "y")
+
+    ts = _ts()
+    result = await _run(args, ts)
+    jp, mp = _write_report(result, ts)
+
+    g = result["grid"]
+    print(f"BLOCK-YOD CALIBRATION: {g['n_scorable']}/{g['n_finals']} scorable finals — "
+          "DIRECTIONAL EVIDENCE, not a regression")
+    n_recon = sum(1 for c in g["cases"] if c["supply_mode"] == "reconstructed")
+    n_stored = sum(1 for c in g["cases"] if c["supply_mode"] == "stored")
+    print(f"  direction supply: {n_stored} stored · {n_recon} reconstructed · "
+          f"{sum(1 for c in g['cases'] if c['supply_mode']=='skip')} skipped")
+    if g["n_scorable"] == 0:
+        print("  no scorable finals (need a parsable discussion section + a suppliable direction).")
+    elif result.get("dry_run"):
+        print(f"  dry-run: {g['total_generations']} generations planned "
+              f"({len(g['efforts'])} efforts × {g['n_scorable']} cases × {g['repeats']} repeats)")
+    else:
+        print(f"  current default={g['current_default']} → recommended={result.get('recommended')}")
+    print(f"  report: {mp}")
+    return 0
+
+
+if __name__ == "__main__":
+    # --self-test is OFFLINE and synchronous (it spins its own short-lived loop for
+    # the overlay probe), so dispatch it BEFORE asyncio.run(main()) — otherwise the
+    # probe's asyncio.run would nest inside main's running loop.
+    if "--self-test" in sys.argv:
+        logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")
+        sys.exit(_self_test())
+    sys.exit(asyncio.run(main()))

scripts/calibrate_effort.py

@@ -19,9 +19,15 @@ WHAT IT MEASURES (per (case, block, effort) cell — the A/B grid):
     • distance                 — normalized composite (lower = closer to Dafna)
   No parallel metric path: it reuses style_distance + learning_loop (G2).
 
-RECOMMENDATION: for each block, the effort with the lowest MEAN composite distance
-across cases (ties → fewer anti-patterns → lower change_percent). Reported next to
-the #204 current default so a regression/improvement is visible.
+RECOMMENDATION (#213 — STYLE-CLEAN, anti-primary): for each block, the effort is
+ranked by `anti_pattern_total` FIRST (the clean Dafna-style-rule-violation count),
+then `golden_ratio_deviation_pp` (structural-style), and `distance` only as a faint
+final tiebreak. `change_percent` is REPORTED-NOT-RANKED: per 07-learning §0.7 it
+mixes style with content-completeness (a case's missing facts the model can't know)
+→ noise for STYLE. The pick carries a `confidence` flag ("clear"/"weak"): a pick
+whose anti-pattern lead over the runner-up is within noise is flagged "weak" so an
+over-claimed pick (the old block-zayin "low") is never shown as confident. Reported
+next to the #204 current default so a regression/improvement is visible.
 
 ⚠️ SAMPLE-SIZE CAVEAT (honored, not hidden): very few cases have an uploaded final
 (draft_final_pairs.final_text non-empty). The report prints n_finals PROMINENTLY and
@@ -55,12 +61,16 @@ from __future__ import annotations
 import argparse
 import asyncio
 import json
+import logging
 import os
 import sys
-from datetime import datetime, timezone
+from datetime import datetime
 from pathlib import Path
+from zoneinfo import ZoneInfo
 from statistics import mean
 
+logger = logging.getLogger(__name__)
+
 REPO_ROOT = Path(__file__).resolve().parent.parent
 sys.path.insert(0, str(REPO_ROOT / "mcp-server" / "src"))
 
@@ -77,28 +87,79 @@ OUT_DIR = REPO_ROOT / "data" / "eval"
 
 # Only the AI blocks that map to a golden-ratio section can be scored against the
 # final's matching section (split_final_by_section). Template blocks (א-ד, יב) are
-# deterministic template-fill — no effort knob. block-yod (discussion) is out of
-# WS5's interim scope but mappable, so it's included when present.
-CALIBRATABLE_BLOCKS = ["block-he", "block-vav", "block-zayin", "block-chet", "block-tet", "block-yod", "block-yod-alef"]
+# deterministic template-fill — no effort knob. block-yod (discussion) is EXCLUDED
+# from the defaults: it requires an approved direction (brainstorm → approve_direction)
+# that calibration cases lack, so block_writer.write_block raises and it is NOT
+# calibratable standalone — and it's out of WS5's interim-draft scope. It stays in
+# VALID_BLOCKS so a user can still force it (--blocks block-yod) at their own risk;
+# the per-cell guard in _run() keeps such a failure non-fatal.
+CALIBRATABLE_BLOCKS = ["block-he", "block-vav", "block-zayin", "block-chet", "block-tet", "block-yod-alef"]
+VALID_BLOCKS = CALIBRATABLE_BLOCKS + ["block-yod"]
 DEFAULT_EFFORTS = ["low", "medium", "high", "xhigh"]
 VALID_EFFORTS = {"low", "medium", "high", "xhigh", "max"}
 
 
 # ── pure helpers (offline-testable) ──────────────────────────────────────────
-def recommend_effort(cells: list[dict]) -> dict | None:
-    """Pick the best effort for ONE block from its scored cells.
+# Style-clean ranking epsilon (#213). Confidence in a pick = how far the best
+# effort's anti_pattern_total beats the 2nd-best. We call a pick "clear" only when
+# that margin clears a floor that is itself the LARGER of:
+#   • an absolute floor (≥0.5 avg anti-patterns — a half-violation per case), and
+#   • a fraction of the spread across efforts (best can't claim a lead that is mere
+#     noise on a block whose efforts barely differ).
+# Below that → "weak" (within noise; nominal pick still reported, but flagged so a
+# within-noise pick is never shown as confident). Documented in the docstring.
+CONFIDENCE_ABS_FLOOR = 0.5       # avg anti-patterns (a half-violation per case)
+CONFIDENCE_SPREAD_FRACTION = 0.20  # of the across-effort anti spread
 
-    cells: [{"effort","distance","anti_pattern_total","change_percent","n"}].
-    Lowest mean composite distance wins; ties broken by fewer anti-patterns,
-    then lower change_percent. Pure → unit-tested in --self-test.
+
+def _anti_confidence(rows: list[dict]) -> tuple[str, float, float]:
+    """Confidence that the anti-primary pick is real, not within-noise.
+
+    Returns (confidence, margin, epsilon) where confidence ∈ {"clear","weak"}.
+    margin = (2nd-best anti) − (best anti); epsilon = the bar margin must clear.
+    A single effort (no rival) is trivially "clear". Pure → unit-tested.
+    """
+    if len(rows) < 2:
+        return "clear", float("inf"), 0.0
+    antis = sorted(r["anti_pattern_total"] for r in rows)
+    best, second = antis[0], antis[1]
+    margin = second - best
+    spread = antis[-1] - antis[0]
+    epsilon = max(CONFIDENCE_ABS_FLOOR, CONFIDENCE_SPREAD_FRACTION * spread)
+    return ("clear" if margin >= epsilon else "weak"), round(margin, 4), round(epsilon, 4)
+
+
+def recommend_effort(cells: list[dict]) -> dict | None:
+    """Pick the STYLE-CLEANEST effort for ONE block from its scored cells (#213).
+
+    cells: [{"effort","distance","anti_pattern_total","change_percent",
+             "golden_ratio_deviation_pp","n"}].
+
+    Ranking key (style-clean, G8 / 07-learning §0.7) — change_percent is EXCLUDED
+    from the ranking (it mixes style with content-completeness; reported only):
+      1. PRIMARY  anti_pattern_total           (↑ fewer Dafna-style violations = better)
+      2. SECONDARY golden_ratio_deviation_pp   (↑ structural-style; None → +inf worst)
+      3. TIEBREAK distance                     (↑ faint final tiebreak only)
+
+    Returns the winning row dict (so callers reading ["effort"] keep working) with
+    two added keys: ``confidence`` ∈ {"clear","weak"} and ``confidence_margin`` —
+    so a within-noise pick is never presented as confident. Pure → unit-tested.
     """
     if not cells:
         return None
     ranked = sorted(
         cells,
-        key=lambda c: (c["distance"], c["anti_pattern_total"], c["change_percent"]),
+        key=lambda c: (
+            c["anti_pattern_total"],
+            c["golden_ratio_deviation_pp"] if c.get("golden_ratio_deviation_pp") is not None else float("inf"),
+            c["distance"],
+        ),
     )
-    return ranked[0]
+    best = dict(ranked[0])
+    conf, margin, _eps = _anti_confidence(cells)
+    best["confidence"] = conf
+    best["confidence_margin"] = margin
+    return best
 
 
 def aggregate_cell(per_run: list[dict]) -> dict:
@@ -124,6 +185,88 @@ def _current_default(block_id: str) -> str | None:
     return cfg.get("effort", DEFAULT_EFFORT)
 
 
+# ── re-rank a saved report (offline — NO LLM, NO DB) ─────────────────────────
+def rerank_saved(saved: dict) -> dict:
+    """Re-apply the CURRENT (#213) recommend_effort to a saved calibration report.
+
+    Reads the per-(block,effort) rows that the run already persisted
+    (distance/anti/change%/ratioΔ/n) and recomputes recommended + confidence — no
+    regeneration, no model, no DB. Lets a past run be re-ranked under the new
+    style-clean key. Returns {"block_id": {old, new, confidence, confidence_margin,
+    current_default, efforts}}. Pure → unit-tested in --self-test.
+    """
+    out: dict[str, dict] = {}
+    for block_id, bd in (saved.get("by_block") or {}).items():
+        rows = bd.get("efforts") or []
+        rec = recommend_effort(rows)
+        out[block_id] = {
+            "section": bd.get("section"),
+            "current_default": bd.get("current_default"),
+            "old_recommended": bd.get("recommended"),
+            "new_recommended": rec["effort"] if rec else None,
+            "confidence": rec["confidence"] if rec else None,
+            "confidence_margin": rec.get("confidence_margin") if rec else None,
+            "efforts": rows,
+        }
+    return out
+
+
+def _render_rerank_md(saved: dict, reranked: dict, src: Path) -> str:
+    g = saved.get("grid", {})
+    finals = g.get("finals", [])
+    lines = [
+        f"# #213 — re-rank של {src.name} (style-clean, anti-primary · OFFLINE)\n",
+        f"> מקור: `{src.name}` · {g.get('n_finals','?')} סופיים "
+        f"({', '.join(finals) or '—'}) · ללא LLM/DB — דירוג-מחדש בלבד.\n",
+        "## old pick → new pick (+confidence)\n",
+        "| block | section | current | old pick | **new pick** | confidence | n |",
+        "|---|---|---|---|---|---|---|",
+    ]
+    for b, rd in reranked.items():
+        old = rd.get("old_recommended") or "—"
+        new = rd.get("new_recommended") or "—"
+        mark = "" if new == old else " ⬅︎"
+        conf = rd.get("confidence")
+        conf_cell = {"clear": "clear", "weak": "⚠️ weak (within-noise)"}.get(conf, "—")
+        n_b = rd["efforts"][0]["n"] if rd.get("efforts") else 0
+        lines.append(
+            f"| {b} | {rd.get('section') or '—'} | {rd.get('current_default') or '—'} | "
+            f"{old} | **{new}**{mark} | {conf_cell} | {n_b} |")
+    lines.append("")
+    for b, rd in reranked.items():
+        if not rd.get("efforts"):
+            continue
+        lines += [f"### {b} ({rd.get('section') or '—'})\n",
+                  "| effort | distance | anti_total | change% (reported) | ratioΔpp | n |",
+                  "|---|---|---|---|---|---|"]
+        for r in rd["efforts"]:
+            star = " ⭐" if r["effort"] == rd.get("new_recommended") else ""
+            ratio = r.get("golden_ratio_deviation_pp")
+            lines.append(
+                f"| {r['effort']}{star} | {r['distance']:.4f} | {r['anti_pattern_total']} | "
+                f"{r['change_percent']} | {ratio if ratio is not None else '—'} | {r['n']} |")
+        lines.append("")
+    lines.append("> דירוג **style-clean** (#213): anti_total ראשי → ratioΔ → distance (tiebreak); "
+                 "**change% מדווח-לא-מדורג** (07-learning §0.7); "
+                 "confidence=⚠️weak ⇒ הובלת-anti < epsilon = max(0.5, 0.20·spread).\n")
+    return "\n".join(lines) + "\n"
+
+
+def _do_rerank(path: Path) -> int:
+    """--rerank entry: load a saved report, re-rank offline, print + write *-reranked.md."""
+    if not path.exists():
+        print(f"rerank: file not found: {path}", file=sys.stderr)
+        return 2
+    saved = json.loads(path.read_text(encoding="utf-8"))
+    reranked = rerank_saved(saved)
+    md = _render_rerank_md(saved, reranked, path)
+    print(md)
+    out = path.with_name(path.stem + "-reranked.md")
+    out.write_text(md, encoding="utf-8")
+    print(f"  reranked report written: {out}", file=sys.stderr)
+    return 0
+
+
 # ── self-test (no DB, no model) ──────────────────────────────────────────────
 def _self_test() -> int:
     ok = True
@@ -171,16 +314,66 @@ def _self_test() -> int:
         {"background", "claims", "discussion", "summary"}))
     chk("split found ≥1 section", len(secs) >= 1)
 
-    # recommend_effort: lowest distance wins; tie → fewer anti-patterns.
+    # recommend_effort (#213, style-clean): anti_pattern_total is PRIMARY — the
+    # effort with the FEWEST style violations wins, even if its distance is higher.
     rec = recommend_effort([
-        {"effort": "low", "distance": 0.40, "anti_pattern_total": 5, "change_percent": 40, "n": 1},
-        {"effort": "high", "distance": 0.20, "anti_pattern_total": 2, "change_percent": 25, "n": 1},
-        {"effort": "xhigh", "distance": 0.20, "anti_pattern_total": 1, "change_percent": 22, "n": 1},
+        {"effort": "low", "distance": 0.20, "anti_pattern_total": 5, "change_percent": 22,
+         "golden_ratio_deviation_pp": 3.0, "n": 1},
+        {"effort": "high", "distance": 0.40, "anti_pattern_total": 2, "change_percent": 40,
+         "golden_ratio_deviation_pp": 5.0, "n": 1},
+        {"effort": "xhigh", "distance": 0.50, "anti_pattern_total": 1, "change_percent": 50,
+         "golden_ratio_deviation_pp": 9.0, "n": 1},
     ])
-    chk("recommend picks lowest distance", rec["distance"] == 0.20)
-    chk("recommend tie → fewer anti-patterns", rec["effort"] == "xhigh")
+    chk("recommend: anti-primary (not distance)", rec["effort"] == "xhigh")
+    chk("recommend: change_percent NOT ranked", rec["distance"] == 0.50)  # worst distance still won
     chk("recommend empty ⇒ None", recommend_effort([]) is None)
 
+    # SECONDARY = golden_ratio_deviation_pp when anti ties; distance only as last tiebreak.
+    rec_tie = recommend_effort([
+        {"effort": "low", "distance": 0.10, "anti_pattern_total": 3, "change_percent": 10,
+         "golden_ratio_deviation_pp": 8.0, "n": 1},
+        {"effort": "high", "distance": 0.90, "anti_pattern_total": 3, "change_percent": 90,
+         "golden_ratio_deviation_pp": 2.0, "n": 1},
+    ])
+    chk("recommend: ratio breaks anti-tie", rec_tie["effort"] == "high")
+    rec_ratio_tie = recommend_effort([
+        {"effort": "low", "distance": 0.30, "anti_pattern_total": 3, "change_percent": 10,
+         "golden_ratio_deviation_pp": 5.0, "n": 1},
+        {"effort": "high", "distance": 0.10, "anti_pattern_total": 3, "change_percent": 90,
+         "golden_ratio_deviation_pp": 5.0, "n": 1},
+    ])
+    chk("recommend: distance breaks final tie", rec_ratio_tie["effort"] == "high")
+    # None ratio is treated as worst (never beats a real deviation on the secondary key).
+    rec_none = recommend_effort([
+        {"effort": "low", "distance": 0.10, "anti_pattern_total": 2, "change_percent": 10,
+         "golden_ratio_deviation_pp": None, "n": 1},
+        {"effort": "high", "distance": 0.90, "anti_pattern_total": 2, "change_percent": 90,
+         "golden_ratio_deviation_pp": 4.0, "n": 1},
+    ])
+    chk("recommend: None ratio is worst", rec_none["effort"] == "high")
+
+    # confidence flag (#213): a clear anti lead ⇒ "clear"; within-noise ⇒ "weak".
+    rec_clear = recommend_effort([
+        {"effort": "low", "distance": 0.5, "anti_pattern_total": 1.0, "change_percent": 50,
+         "golden_ratio_deviation_pp": 5.0, "n": 5},
+        {"effort": "high", "distance": 0.5, "anti_pattern_total": 4.0, "change_percent": 50,
+         "golden_ratio_deviation_pp": 5.0, "n": 5},
+    ])
+    chk("confidence: big anti lead ⇒ clear", rec_clear["confidence"] == "clear")
+    rec_weak = recommend_effort([
+        {"effort": "low", "distance": 0.4, "anti_pattern_total": 3.6, "change_percent": 93,
+         "golden_ratio_deviation_pp": 6.7, "n": 5},
+        {"effort": "xhigh", "distance": 0.5, "anti_pattern_total": 3.6, "change_percent": 102,
+         "golden_ratio_deviation_pp": 6.6, "n": 5},
+        {"effort": "high", "distance": 0.53, "anti_pattern_total": 3.8, "change_percent": 104,
+         "golden_ratio_deviation_pp": 5.7, "n": 5},
+    ])
+    chk("confidence: anti tie ⇒ weak", rec_weak["confidence"] == "weak")
+    rec_single = recommend_effort([
+        {"effort": "low", "distance": 0.4, "anti_pattern_total": 2.0, "change_percent": 50,
+         "golden_ratio_deviation_pp": 5.0, "n": 5}])
+    chk("confidence: single effort ⇒ clear", rec_single["confidence"] == "clear")
+
     # aggregate_cell: means + ratio drops Nones, keeps n.
     agg = aggregate_cell([
         {"distance": 0.2, "anti_pattern_total": 2, "change_percent": 20, "golden_ratio_deviation_pp": 3.0},
@@ -190,6 +383,23 @@ def _self_test() -> int:
     chk("aggregate n counted", agg["n"] == 2)
     chk("aggregate ratio skips None", agg["golden_ratio_deviation_pp"] == 3.0)
 
+    # rerank_saved: re-applies the new key to a saved report's stored rows, no I/O.
+    # Mimics the real 8-final block-vav rows: low has fewest anti by a clear margin.
+    saved = {"grid": {"n_finals": 7, "finals": ["x"]}, "by_block": {
+        "block-vav": {"section": "background", "current_default": "medium",
+                      "recommended": "low", "efforts": [
+            {"effort": "low", "distance": 0.4454, "anti_pattern_total": 1.57, "change_percent": 87.39, "golden_ratio_deviation_pp": 6.33, "n": 7},
+            {"effort": "medium", "distance": 0.4749, "anti_pattern_total": 2.29, "change_percent": 87.76, "golden_ratio_deviation_pp": 6.37, "n": 7},
+            {"effort": "high", "distance": 0.5381, "anti_pattern_total": 4.14, "change_percent": 90.69, "golden_ratio_deviation_pp": 5.99, "n": 7},
+            {"effort": "xhigh", "distance": 0.5413, "anti_pattern_total": 4.57, "change_percent": 89.53, "golden_ratio_deviation_pp": 6.41, "n": 7},
+        ]},
+    }}
+    rr = rerank_saved(saved)
+    chk("rerank: block-vav stays low", rr["block-vav"]["new_recommended"] == "low")
+    chk("rerank: block-vav clear (anti lead)", rr["block-vav"]["confidence"] == "clear")
+    chk("rerank: keeps old pick for diff", rr["block-vav"]["old_recommended"] == "low")
+    chk("rerank: renders markdown", "new pick" in _render_rerank_md(saved, rr, Path("x.json")))
+
     print("ALL PASS" if ok else "*** FAILURES ***")
     return 0 if ok else 1
 
@@ -228,7 +438,7 @@ async def _score_cell(case_id, block_id: str, effort: str, final_section: str,
     return agg
 
 
-async def _run(args) -> dict:
+async def _run(args, ts: str) -> dict:
     from uuid import UUID
     from legal_mcp.services import db
     from legal_mcp.services.lessons import canonical_outcome
@@ -279,10 +489,21 @@ async def _run(args) -> dict:
             final_section = c["sections"][section]
             case_cells = []
             for effort in efforts:
+                # Per-cell guard: ANY failure of a single (case, block, effort) cell
+                # — e.g. block-yod raising "ללא כיוון מאושר", or a transient rate-limit
+                # mid-run — is logged and skipped, NOT allowed to kill the whole run
+                # (INV-G8 eval-harness robustness). Partial results still persist below.
+                try:
                     cell = await _score_cell(
                         UUID(c["case_id"]), block_id, effort, final_section,
                         c["final_total_words"], c["outcome"], args.repeats,
                     )
+                except Exception as exc:  # noqa: BLE001 — harness must survive any cell failure
+                    logger.warning(
+                        "calibration cell skipped: case=%s block=%s effort=%s — %s",
+                        c["case_number"], block_id, effort, exc,
+                    )
+                    continue
                 per_effort_runs[effort].append(cell)
                 case_cells.append({k: cell[k] for k in
                                    ("effort", "distance", "anti_pattern_total",
@@ -309,15 +530,35 @@ async def _run(args) -> dict:
             "section": section,
             "current_default": _current_default(block_id),
             "recommended": rec["effort"] if rec else None,
+            "confidence": rec["confidence"] if rec else None,
+            "confidence_margin": rec.get("confidence_margin") if rec else None,
             "efforts": effort_rows,
             "per_case": per_case,
         }
 
+        # Incremental persistence (INV-G8): flush the report after every completed
+        # block so a crash later in the grid never loses blocks already generated.
+        # Blocks not yet done are simply absent from by_block; _write_report tolerates
+        # partial results. main() does the final flush once the loop finishes.
+        try:
+            _write_report({"dry_run": False, "grid": grid_summary, "by_block": by_block}, ts)
+        except Exception as exc:  # noqa: BLE001 — a write hiccup must not abort the run
+            logger.warning("incremental report write failed after block=%s — %s", block_id, exc)
+
     return {"dry_run": False, "grid": grid_summary, "by_block": by_block}
 
 
+IL_TZ = ZoneInfo("Asia/Jerusalem")
+
+
 def _ts() -> str:
-    return datetime.now(timezone.utc).strftime("%Y%m%dT%H%M%SZ")
+    """Filename-safe report stamp in Israel time (INV-UI9: human-facing display).
+
+    This script runs on the host, which is UTC, so a process-clock stamp would
+    mislead the chair. We pin Asia/Jerusalem and suffix ``IL`` (instead of a bare
+    ``Z``) so the displayed/filename time is unambiguously Israel-local.
+    """
+    return datetime.now(IL_TZ).strftime("%Y%m%dT%H%M%S-IL")
 
 
 def _write_report(result: dict, ts: str) -> tuple[Path, Path]:
@@ -328,8 +569,10 @@ def _write_report(result: dict, ts: str) -> tuple[Path, Path]:
 
     g = result["grid"]
     n = g["n_finals"]
+    il_now = datetime.now(IL_TZ).strftime("%Y-%m-%d %H:%M")
     lines = [
         f"# #208 — כיול model×effort מול הסופיים — {ts}\n",
+        f"> נוצר: {il_now} (שעון ישראל · Asia/Jerusalem)\n",
         f"> ⚠️ **גודל-מדגם: {n} סופיים** ({', '.join(g['finals']) or '—'}). "
         "זוהי **עדות-כיוון, לא רגרסיה** — מעט תיקים בעלי סופי-עלוי. "
         "ההמלצה אדוויזורית; ההכרעה בידי היו\"ר/המפעיל.\n",
@@ -346,13 +589,18 @@ def _write_report(result: dict, ts: str) -> tuple[Path, Path]:
         mp.write_text("\n".join(lines) + "\n", encoding="utf-8")
         return jp, mp
 
-    lines += ["## המלצה per-בלוק (distance נמוך = קרוב יותר לדפנה)\n",
-              "| block | section | current | **recommended** | n |", "|---|---|---|---|---|"]
+    lines += ["## המלצה per-בלוק (style-clean: anti_pattern_total ראשי — #213)\n",
+              "| block | section | current | **recommended** | confidence | n |",
+              "|---|---|---|---|---|---|"]
     for b, bd in result["by_block"].items():
         rec = bd.get("recommended") or "—"
         mark = "" if rec == bd.get("current_default") else " ⬅︎"
+        conf = bd.get("confidence")
+        conf_cell = {"clear": "clear", "weak": "⚠️ weak (within-noise)"}.get(conf, "—")
         n_b = bd["efforts"][0]["n"] if bd.get("efforts") else 0
-        lines.append(f"| {b} | {bd.get('section','')} | {bd.get('current_default') or '—'} | **{rec}**{mark} | {n_b} |")
+        lines.append(
+            f"| {b} | {bd.get('section','')} | {bd.get('current_default') or '—'} | "
+            f"**{rec}**{mark} | {conf_cell} | {n_b} |")
     lines.append("")
     for b, bd in result["by_block"].items():
         lines += [f"### {b} ({bd.get('section','')})\n",
@@ -365,8 +613,10 @@ def _write_report(result: dict, ts: str) -> tuple[Path, Path]:
                 f"| {r['effort']}{star} | {r['distance']:.4f} | {r['anti_pattern_total']} | "
                 f"{r['change_percent']} | {ratio if ratio is not None else '—'} | {r['n']} |")
         lines.append("")
-    lines.append("> change% מערבב סגנון עם שלמות-תוכן (07-learning §0.7); "
-                 "anti_total הוא הסיגנל הנקי-יותר לסגנון.\n")
+    lines.append("> דירוג-ההמלצה **style-clean** (#213): anti_total ראשי → ratioΔ → distance (tiebreak). "
+                 "**change% מדווח-לא-מדורג** — מערבב סגנון עם שלמות-תוכן (07-learning §0.7), "
+                 "anti_total הוא הסיגנל הנקי-לסגנון. confidence=⚠️weak ⇒ הבחירה בתוך-הרעש "
+                 "(ההובלה ב-anti קטנה מ-epsilon = max(0.5, 0.20·spread)).\n")
     mp.write_text("\n".join(lines) + "\n", encoding="utf-8")
     return jp, mp
 
@@ -374,6 +624,8 @@ def _write_report(result: dict, ts: str) -> tuple[Path, Path]:
 async def main() -> int:
     ap = argparse.ArgumentParser(description="#208 model/effort calibration harness")
     ap.add_argument("--self-test", action="store_true", help="offline measurement-logic proof (no DB/CLI)")
+    ap.add_argument("--rerank", metavar="REPORT.json", default=None,
+                    help="re-rank a saved calibration JSON under the current style-clean key (no LLM/DB)")
     ap.add_argument("--dry-run", action="store_true", help="plan the A/B grid over existing finals, no model calls")
     ap.add_argument("--efforts", default=",".join(DEFAULT_EFFORTS),
                     help=f"comma effort grid (default {','.join(DEFAULT_EFFORTS)})")
@@ -383,22 +635,27 @@ async def main() -> int:
     ap.add_argument("--repeats", type=int, default=1, help="generations per cell (avg out gen noise)")
     args = ap.parse_args()
 
+    logging.basicConfig(level=logging.INFO, format="%(asctime)s %(levelname)s %(message)s")
+
     if args.self_test:
         return _self_test()
 
+    if args.rerank:
+        return _do_rerank(Path(args.rerank))
+
     args.efforts = [e.strip() for e in args.efforts.split(",") if e.strip()]
     bad = [e for e in args.efforts if e not in VALID_EFFORTS]
     if bad:
         print(f"invalid effort(s): {bad}. valid: {sorted(VALID_EFFORTS)}", file=sys.stderr)
         return 2
     args.blocks = [b.strip() for b in args.blocks.split(",") if b.strip()]
-    bad_b = [b for b in args.blocks if b not in CALIBRATABLE_BLOCKS]
+    bad_b = [b for b in args.blocks if b not in VALID_BLOCKS]
     if bad_b:
-        print(f"non-calibratable block(s): {bad_b}. valid: {CALIBRATABLE_BLOCKS}", file=sys.stderr)
+        print(f"non-calibratable block(s): {bad_b}. valid: {VALID_BLOCKS}", file=sys.stderr)
         return 2
 
-    result = await _run(args)
     ts = _ts()
+    result = await _run(args, ts)
     jp, mp = _write_report(result, ts)
 
     g = result["grid"]
@@ -410,7 +667,8 @@ async def main() -> int:
     else:
         for b, bd in result["by_block"].items():
             print(f"  {b:16} current={bd.get('current_default') or '—':6} "
-                  f"→ recommended={bd.get('recommended') or '—'}")
+                  f"→ recommended={bd.get('recommended') or '—':6} "
+                  f"[{bd.get('confidence') or '—'}]")
     print(f"  report: {mp}")
     return 0
 

web-ui/src/app/approvals/page.tsx

@@ -9,6 +9,7 @@ import {
   type ApprovalCategory,
   type ApprovalSeverity,
 } from "@/lib/api/chair";
+import { formatDateLong } from "@/lib/format-date";
 
 /**
  * מרכז אישורים — דפנה (INV-G10).
@@ -35,18 +36,7 @@ const SEVERITY_RAIL: Record<ApprovalSeverity, string> = {
   ok: "border-s-success",
 };
 
-function formatDate(iso?: string | null): string {
-  if (!iso) return "";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL", {
-      day: "numeric",
-      month: "long",
-      year: "numeric",
-    });
-  } catch {
-    return "";
-  }
-}
+const formatDate = (iso?: string | null) => formatDateLong(iso);
 
 function ApprovalCard({ cat }: { cat: ApprovalCategory }) {
   const cleared = cat.count === 0;

web-ui/src/app/archive/page.tsx

@@ -28,19 +28,7 @@ import {
 import { useCases, useRestoreCase, type Case } from "@/lib/api/cases";
 import { subtypeOf } from "@/components/cases/appeal-type-bars";
 import { APPEAL_SUBTYPE_LABELS, type AppealSubtype } from "@/lib/practice-area";
-
-function formatDate(iso?: string | null) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL", {
-      day: "2-digit",
-      month: "2-digit",
-      year: "numeric",
-    });
-  } catch {
-    return iso;
-  }
-}
+import { formatDate, getIsraelYear } from "@/lib/format-date";
 
 // type chip styling per mockup 05 (.t-lic / .t-bet / .t-comp)
 const TYPE_CHIP: Record<string, string> = {
@@ -177,8 +165,8 @@ export default function ArchivePage() {
     const set = new Set<string>();
     for (const c of rows) {
       if (!c.archived_at) continue;
-      const y = new Date(c.archived_at).getFullYear();
-      if (!Number.isNaN(y)) set.add(String(y));
+      const y = getIsraelYear(c.archived_at);
+      if (y !== null) set.add(String(y));
     }
     return [...set].sort((a, b) => Number(b) - Number(a));
   }, [rows]);
@@ -209,7 +197,7 @@ export default function ArchivePage() {
     if (yearFilter !== "all") {
       all = all.filter((r) => {
         const iso = r.original.archived_at;
-        return iso != null && String(new Date(iso).getFullYear()) === yearFilter;
+        return iso != null && String(getIsraelYear(iso)) === yearFilter;
       });
     }
     return all;

web-ui/src/app/cases/[caseNumber]/compose/page.tsx

@@ -1,506 +0,0 @@
-"use client";
-
-import { use, useRef, useState } from "react";
-import Link from "next/link";
-import { FileText } from "lucide-react";
-import { AppShell } from "@/components/app-shell";
-import { Card, CardContent } from "@/components/ui/card";
-import { Button } from "@/components/ui/button";
-import { Skeleton } from "@/components/ui/skeleton";
-import { SubsectionCard } from "@/components/compose/subsection-card";
-import { PrecedentsSection } from "@/components/compose/precedents-section";
-import { CitationVerificationPanel } from "@/components/compose/citation-verification-panel";
-import { DecisionBlocksPanel } from "@/components/cases/decision-blocks-panel";
-import { Markdown } from "@/components/ui/markdown";
-import { Tabs, TabsContent, TabsList, TabsTrigger } from "@/components/ui/tabs";
-import { useCase, type CaseStatus } from "@/lib/api/cases";
-import {
-  useCaseLearningStatus,
-  type CaseLearningStatus,
-} from "@/lib/api/learning";
-import { useResearchAnalysis } from "@/lib/api/research";
-import { useCasePrecedents } from "@/lib/api/precedents";
-import { APPEAL_SUBTYPES } from "@/lib/practice-area";
-import { DOC_TYPE_LABELS, type DocType } from "@/lib/doc-types";
-
-// ── Case-status → Hebrew label + tone (mockup 03 status chip) ────────────────
-const STATUS_CHIP: Record<string, { label: string; cls: string }> = {
-  new: { label: "חדש", cls: "bg-rule-soft text-ink-muted border-rule" },
-  processing: { label: "בעיבוד", cls: "bg-info-bg text-info border-info/30" },
-  documents_ready: { label: "מסמכים מוכנים", cls: "bg-info-bg text-info border-info/30" },
-  outcome_set: { label: "תוצאה נקבעה", cls: "bg-info-bg text-info border-info/30" },
-  direction_approved: { label: "כיוון אושר", cls: "bg-info-bg text-info border-info/30" },
-  qa_review: { label: "בדיקת-איכות", cls: "bg-gold-wash text-gold-deep border-gold/40" },
-  drafted: { label: "טיוטה", cls: "bg-gold-wash text-gold-deep border-gold/40" },
-  exported: { label: "יוצא", cls: "bg-success-bg text-success border-success/40" },
-  reviewed: { label: "נסקר", cls: "bg-success-bg text-success border-success/40" },
-  final: { label: "סופי", cls: "bg-success-bg text-success border-success/40" },
-};
-
-function StatusChip({ status }: { status?: CaseStatus }) {
-  const c = (status && STATUS_CHIP[status]) || {
-    label: "בעריכה",
-    cls: "bg-info-bg text-info border-info/30",
-  };
-  return (
-    <span
-      className={`rounded-full text-[0.78rem] font-semibold px-3 py-0.5 border ${c.cls}`}
-    >
-      {c.label}
-    </span>
-  );
-}
-
-function subtypeLabel(subtype?: string | null): string | null {
-  if (!subtype) return null;
-  return APPEAL_SUBTYPES.find((s) => s.value === subtype)?.label ?? null;
-}
-
-// ── Staged-pipeline indicator text (mockup 03) — derived from the live
-// learning-status, same source as the drafts-panel LearningStatusBadges. ──────
-function voiceLearningText(s?: CaseLearningStatus): string {
-  if (!s?.final_uploaded) return "ממתין להעלאת הסופי";
-  const v = s.voice_learning;
-  if (v.outcome === "succeeded") {
-    const bits = [`${v.lessons_count} לקחים הופקו`];
-    if (v.lessons_proposed > 0) bits.push(`${v.lessons_proposed} הוצעו לאישור`);
-    return `✓ הושלם · ${bits.join(" · ")}`;
-  }
-  if (v.outcome === "failed") return v.error ? `✗ נכשל — ${v.error}` : "✗ נכשל";
-  return "ממתין להרצה";
-}
-
-function halachaExtractionText(s?: CaseLearningStatus): string {
-  if (!s?.final_uploaded) return "ממתין להעלאת הסופי";
-  const h = s.halacha_extraction;
-  if (!h.enrolled_in_corpus)
-    return h.not_enrolled_reason ?? "לא נכנס לקורפוס-הפסיקה";
-  switch (h.status) {
-    case "completed":
-      return `✓ הושלם · חולצו ${h.halachot_count} · ${h.approved} אושרו · ${h.rejected} נדחו`;
-    case "processing":
-      return "רץ עכשיו…";
-    case "pending":
-    case "busy":
-      return "בתור";
-    case "partial":
-      return `חלקי · חולצו ${h.halachot_count}`;
-    case "failed":
-    case "extraction_failed":
-      return "✗ נכשל";
-    case "no_chunks":
-      return "אין טקסט לחילוץ";
-    default:
-      return "ממתין להרצה";
-  }
-}
-
-function ProseSection({ title, content }: { title: string; content?: string }) {
-  if (!content?.trim()) return null;
-  return (
-    <section className="space-y-2">
-      <h3 className="text-[0.78rem] uppercase tracking-[0.08em] text-gold-deep font-semibold">
-        {title}
-      </h3>
-      <Markdown content={content.trim()} />
-    </section>
-  );
-}
-
-// ── "השלמה והעברה" rail card — DOCX export, upload, download (all real) ──────
-function FinishRail({
-  caseNumber,
-  hasAnalysis,
-  onUploaded,
-}: {
-  caseNumber: string;
-  hasAnalysis: boolean;
-  onUploaded: () => void;
-}) {
-  const fileRef = useRef<HTMLInputElement>(null);
-  const [uploading, setUploading] = useState(false);
-  const [uploadMsg, setUploadMsg] = useState<{ ok: boolean; text: string } | null>(null);
-  const learning = useCaseLearningStatus(caseNumber);
-
-  async function handleUpload(file: File) {
-    setUploading(true);
-    setUploadMsg(null);
-    try {
-      const form = new FormData();
-      form.append("file", file);
-      const res = await fetch(`/api/cases/${caseNumber}/research/analysis/upload`, {
-        method: "PUT",
-        body: form,
-      });
-      const data = await res.json();
-      if (!res.ok) {
-        setUploadMsg({ ok: false, text: data.detail || "שגיאה בהעלאה" });
-        return;
-      }
-      setUploadMsg({
-        ok: true,
-        text: `הקובץ הועלה — ${data.sections.threshold_claims} טענות סף, ${data.sections.issues} סוגיות`,
-      });
-      onUploaded();
-    } catch {
-      setUploadMsg({ ok: false, text: "שגיאת רשת" });
-    } finally {
-      setUploading(false);
-      if (fileRef.current) fileRef.current.value = "";
-    }
-  }
-
-  return (
-    <Card className="bg-surface border-rule shadow-sm">
-      <CardContent className="px-4 py-4">
-        <h3 className="text-navy text-[0.9rem] font-semibold mb-3">השלמה והעברה</h3>
-
-        <input
-          ref={fileRef}
-          type="file"
-          accept=".md"
-          className="hidden"
-          onChange={(e) => {
-            const f = e.target.files?.[0];
-            if (f) handleUpload(f);
-          }}
-        />
-
-        <div className="space-y-2">
-          {hasAnalysis && (
-            <Button
-              variant="outline"
-              className="w-full justify-center"
-              onClick={() => {
-                const a = document.createElement("a");
-                a.href = `/api/cases/${caseNumber}/research/analysis/export-docx`;
-                a.click();
-              }}
-            >
-              ייצוא DOCX
-            </Button>
-          )}
-          <Button
-            className="w-full justify-center bg-gold text-white hover:bg-gold-deep"
-            disabled={uploading}
-            onClick={() => fileRef.current?.click()}
-          >
-            {uploading ? "מעלה…" : "העלאת ניתוח מעודכן"}
-          </Button>
-          {hasAnalysis && (
-            <Button
-              variant="outline"
-              className="w-full justify-center"
-              onClick={() => {
-                const a = document.createElement("a");
-                a.href = `/api/cases/${caseNumber}/research/analysis/download`;
-                a.download = `analysis-${caseNumber}.md`;
-                a.click();
-              }}
-            >
-              הורד ניתוח (MD)
-            </Button>
-          )}
-        </div>
-
-        {uploadMsg && (
-          <p className={`text-xs mt-2 ${uploadMsg.ok ? "text-success" : "text-danger"}`}>
-            {uploadMsg.text}
-          </p>
-        )}
-
-        {/* mockup 03: stage indicators — informational pointers, not actions */}
-        <div className="mt-3 space-y-0">
-          <div className="text-[0.78rem] text-ink-muted pt-2 border-t border-rule-soft">
-            <b className="text-navy">הרץ למידת-קול</b> — {voiceLearningText(learning.data)}
-          </div>
-          <div className="text-[0.78rem] text-ink-muted pt-2 mt-2 border-t border-rule-soft">
-            <b className="text-navy">הרץ אימות-הלכות</b> — {halachaExtractionText(learning.data)}
-          </div>
-        </div>
-
-        <Button asChild variant="ghost" className="w-full justify-center mt-3 text-ink-muted">
-          <Link href={`/cases/${caseNumber}`}>חזרה לתיק</Link>
-        </Button>
-      </CardContent>
-    </Card>
-  );
-}
-
-export default function ComposePage({
-  params,
-}: {
-  params: Promise<{ caseNumber: string }>;
-}) {
-  const { caseNumber } = use(params);
-  const caseQuery = useCase(caseNumber);
-  const analysis = useResearchAnalysis(caseNumber);
-  const precedentsQuery = useCasePrecedents(caseNumber);
-
-  /* Partition the flat list into scopes so each child renders its own slice
-   * without re-fetching. Done once at the page level. */
-  const allPrecedents = precedentsQuery.data ?? [];
-  const caseLevelPrecedents = allPrecedents.filter((p) => p.section_id === null);
-  const precedentsBySection = new Map<string, typeof allPrecedents>();
-  for (const p of allPrecedents) {
-    if (p.section_id) {
-      const existing = precedentsBySection.get(p.section_id) ?? [];
-      existing.push(p);
-      precedentsBySection.set(p.section_id, existing);
-    }
-  }
-  const practiceArea = caseQuery.data?.practice_area ?? null;
-  const subtype = subtypeLabel(caseQuery.data?.appeal_subtype);
-  const parties = (() => {
-    const c = caseQuery.data;
-    if (!c) return null;
-    const app = c.appellants?.length ? c.appellants.join(", ") : null;
-    const resp = c.respondents?.length ? c.respondents.join(", ") : null;
-    const out: string[] = [];
-    if (app) out.push(`עוררים: ${app}`);
-    if (resp) out.push(`משיבה: ${resp}`);
-    return out.length ? out.join(" · ") : c.title || null;
-  })();
-  const documents = caseQuery.data?.documents ?? [];
-
-  const isNotFound =
-    analysis.error instanceof Error &&
-    /404|לא נמצא|טרם בוצע/.test(analysis.error.message);
-
-  return (
-    <AppShell>
-      {/* ── Case header band (mockup 03) — parchment strip, full-bleed to the
-          AppShell <main> edges (which pads px-10 py-10) ── */}
-      <div className="-mx-10 -mt-10 mb-6 border-b border-rule bg-parchment px-10 py-5">
-        <nav className="text-[0.78rem] text-ink-muted flex items-center gap-2 mb-2">
-          <Link href="/" className="hover:text-gold-deep">בית</Link>
-          <span aria-hidden>·</span>
-          <Link href={`/cases/${caseNumber}`} className="hover:text-gold-deep">
-            ערר {caseNumber}
-          </Link>
-          <span aria-hidden>·</span>
-          <span className="text-navy">עורך החלטה</span>
-        </nav>
-        <Link
-          href={`/cases/${caseNumber}`}
-          className="inline-flex items-center gap-1.5 mb-3 rounded-md border border-rule bg-surface px-3 py-1.5 text-[0.82rem] font-semibold text-gold-deep hover:bg-gold-wash hover:text-navy transition-colors"
-        >
-          <span aria-hidden>→</span> חזרה לדף התיק
-        </Link>
-        <div className="flex items-center gap-3 flex-wrap">
-          <h1 className="text-navy text-2xl font-bold mb-0">ערר {caseNumber}</h1>
-          <StatusChip status={caseQuery.data?.status} />
-          {subtype && (
-            <span className="rounded-full text-[0.78rem] font-semibold px-3 py-0.5 border border-rule bg-gold-wash text-gold-deep">
-              {subtype}
-            </span>
-          )}
-          {/* INV-G10: source-of-truth pill — the blocks are the canonical text */}
-          <span className="ms-auto rounded-lg text-[0.8rem] font-semibold px-3.5 py-1.5 border border-gold bg-gold-wash text-gold-deep">
-            מקור-אמת: בלוקים
-          </span>
-        </div>
-        {parties && <p className="text-ink-soft text-sm mt-2">{parties}</p>}
-      </div>
-
-      {caseQuery.isPending ? (
-        <Card className="bg-surface border-rule shadow-sm">
-          <CardContent className="px-6 py-5 space-y-3">
-            <Skeleton className="h-6 w-48" />
-            <Skeleton className="h-4 w-96" />
-            <Skeleton className="h-4 w-80" />
-            <Skeleton className="h-32 w-full" />
-          </CardContent>
-        </Card>
-      ) : (
-        /* ── Two-column workspace: tabbed main editor + 320px side rail ──── */
-        <div className="grid gap-6 lg:grid-cols-[1fr_320px] items-start">
-          {/* MAIN — block editor (default) + chair positions, as tabs (mockup 03) */}
-          <div className="min-w-0">
-            <Tabs defaultValue="blocks" dir="rtl">
-              <TabsList className="bg-rule-soft/60">
-                <TabsTrigger value="blocks">עורך הבלוקים</TabsTrigger>
-                <TabsTrigger value="positions">עמדות וטענות</TabsTrigger>
-                <TabsTrigger value="verify">אימות פסיקה</TabsTrigger>
-              </TabsList>
-
-              {/* Tab 1 — the 12-block decision editor (reused DecisionBlocksPanel) */}
-              <TabsContent value="blocks" className="mt-5">
-                <DecisionBlocksPanel caseNumber={caseNumber} />
-              </TabsContent>
-
-              {/* Tab 3 — citation verification: per-argument support + verify gate (#154) */}
-              <TabsContent value="verify" className="mt-5">
-                <CitationVerificationPanel caseNumber={caseNumber} />
-              </TabsContent>
-
-              {/* Tab 2 — chair positions on the analyst's threshold-claims + issues */}
-              <TabsContent value="positions" className="mt-5">
-                {analysis.isPending ? (
-                  <Card className="bg-surface border-rule shadow-sm">
-                    <CardContent className="px-6 py-5 space-y-3">
-                      <Skeleton className="h-6 w-48" />
-                      <Skeleton className="h-4 w-96" />
-                      <Skeleton className="h-32 w-full" />
-                    </CardContent>
-                  </Card>
-                ) : isNotFound ? (
-                  <Card className="bg-surface border-rule shadow-sm">
-                    <CardContent className="px-6 py-12 text-center space-y-3">
-                      <div className="text-gold text-3xl" aria-hidden>❦</div>
-                      <h2 className="text-navy text-lg mb-0">
-                        טרם בוצע ניתוח משפטי לתיק זה
-                      </h2>
-                      <p className="text-ink-muted text-sm max-w-md mx-auto">
-                        לאחר שקובץ <code>analysis-and-research.md</code> ייווצר, תוכלי
-                        לערוך כאן את עמדת הוועדה לכל טענת סף וסוגיה.
-                      </p>
-                    </CardContent>
-                  </Card>
-                ) : analysis.error ? (
-                  <Card className="bg-danger-bg border-danger/40">
-                    <CardContent className="px-6 py-5 text-center">
-                      <p className="text-danger">{analysis.error.message}</p>
-                    </CardContent>
-                  </Card>
-                ) : analysis.data ? (
-                  <div className="space-y-6">
-                    {/* Threshold claims */}
-            {analysis.data.threshold_claims &&
-              analysis.data.threshold_claims.length > 0 && (
-                <div className="space-y-3">
-                  <div className="flex items-center gap-2">
-                    <h2 className="text-navy text-lg font-semibold mb-0">טענות סף</h2>
-                    <span className="text-[0.72rem] rounded-full bg-gold-wash text-gold-deep px-2 py-0.5 border border-gold/40 tabular-nums">
-                      {analysis.data.threshold_claims.length}
-                    </span>
-                  </div>
-                  <div className="space-y-2.5">
-                    {analysis.data.threshold_claims.map((tc) => (
-                      <SubsectionCard
-                        key={tc.id}
-                        caseNumber={caseNumber}
-                        item={tc}
-                        precedents={precedentsBySection.get(tc.id) ?? []}
-                        practiceArea={practiceArea}
-                      />
-                    ))}
-                  </div>
-                </div>
-              )}
-
-            {/* Issues */}
-            {analysis.data.issues && analysis.data.issues.length > 0 && (
-              <div className="space-y-3">
-                <div className="flex items-center gap-2">
-                  <h2 className="text-navy text-lg font-semibold mb-0">סוגיות להכרעה</h2>
-                  <span className="text-[0.72rem] rounded-full bg-gold-wash text-gold-deep px-2 py-0.5 border border-gold/40 tabular-nums">
-                    {analysis.data.issues.length}
-                  </span>
-                </div>
-                <div className="space-y-2.5">
-                  {analysis.data.issues.map((iss) => (
-                    <SubsectionCard
-                      key={iss.id}
-                      caseNumber={caseNumber}
-                      item={iss}
-                      precedents={precedentsBySection.get(iss.id) ?? []}
-                      practiceArea={practiceArea}
-                    />
-                  ))}
-                </div>
-              </div>
-            )}
-
-            {!analysis.data.threshold_claims?.length &&
-              !analysis.data.issues?.length && (
-                <Card className="bg-surface border-rule">
-                  <CardContent className="px-6 py-10 text-center text-ink-muted">
-                    לא נמצאו טענות סף או סוגיות בניתוח זה.
-                  </CardContent>
-                </Card>
-              )}
-
-            {/* Background prose — supporting context after the decision points */}
-            <Card className="bg-surface border-rule shadow-sm">
-              <CardContent className="px-6 py-5 space-y-5">
-                <h2 className="text-navy text-lg font-semibold mb-0">רקע לניתוח</h2>
-                <ProseSection title="צד מיוצג" content={analysis.data.represented_party} />
-                <ProseSection title="רקע דיוני" content={analysis.data.procedural_background} />
-                <ProseSection title="עובדות מוסכמות" content={analysis.data.agreed_facts} />
-                <ProseSection title="עובדות במחלוקת" content={analysis.data.disputed_facts} />
-              </CardContent>
-            </Card>
-
-            {analysis.data.conclusions?.trim() && (
-              <Card className="bg-gold-wash border-gold/40 shadow-sm">
-                <CardContent className="px-6 py-5 space-y-3">
-                  <h2 className="text-gold-deep text-lg font-semibold mb-0">מסקנות</h2>
-                  <Markdown content={analysis.data.conclusions.trim()} />
-                </CardContent>
-              </Card>
-            )}
-                  </div>
-                ) : null}
-              </TabsContent>
-            </Tabs>
-          </div>
-
-          {/* SIDE RAIL — documents · attached precedents · finish-and-transfer */}
-          <aside className="space-y-4 lg:sticky lg:top-4">
-            {/* מסמכי התיק */}
-            <Card className="bg-surface border-rule shadow-sm">
-              <CardContent className="px-4 py-4">
-                <h3 className="text-navy text-[0.9rem] font-semibold mb-2">מסמכי התיק</h3>
-                {documents.length === 0 ? (
-                  <p className="text-[0.78rem] text-ink-muted">אין מסמכים מצורפים</p>
-                ) : (
-                  <ul>
-                    {documents.map((d) => (
-                      <li
-                        key={d.id}
-                        className="flex items-center gap-2 text-[0.82rem] text-ink-soft py-1.5 border-b border-rule-soft last:border-0"
-                      >
-                        <FileText className="w-3.5 h-3.5 text-ink-muted shrink-0" aria-hidden />
-                        <span className="truncate flex-1" title={d.title}>
-                          {d.title || "מסמך"}
-                        </span>
-                        <span className="rounded bg-rule-soft text-ink-muted text-[0.68rem] px-1.5 py-0.5 shrink-0 whitespace-nowrap">
-                          {DOC_TYPE_LABELS[d.doc_type as DocType] ?? d.doc_type}
-                        </span>
-                      </li>
-                    ))}
-                  </ul>
-                )}
-              </CardContent>
-            </Card>
-
-            {/* פסיקה מצורפת (case-level) */}
-            <Card className="bg-surface border-rule shadow-sm">
-              <CardContent className="px-4 py-4">
-                <h3 className="text-navy text-[0.9rem] font-semibold mb-1">פסיקה מצורפת</h3>
-                <p className="text-[0.72rem] text-ink-muted mb-3">
-                  ציטוטים התומכים בעמדה באופן רוחבי — ישולבו בפתיחת בלוק י (דיון).
-                </p>
-                <PrecedentsSection
-                  caseNumber={caseNumber}
-                  sectionId={null}
-                  precedents={caseLevelPrecedents}
-                  practiceArea={practiceArea}
-                  emptyHelperText="עדיין לא צורפה פסיקה כללית לתיק"
-                />
-              </CardContent>
-            </Card>
-
-            {/* השלמה והעברה */}
-            <FinishRail
-              caseNumber={caseNumber}
-              hasAnalysis={!!analysis.data}
-              onUploaded={() => analysis.refetch()}
-            />
-          </aside>
-        </div>
-      )}
-    </AppShell>
-  );
-}

web-ui/src/app/cases/[caseNumber]/page.tsx

@@ -1,33 +1,34 @@
 "use client";
 
-import { use } from "react";
+import { use, useState } from "react";
 import Link from "next/link";
 import { AppShell } from "@/components/app-shell";
 import { Card, CardContent } from "@/components/ui/card";
 import { Tabs, TabsContent, TabsList, TabsTrigger } from "@/components/ui/tabs";
+import {
+  Accordion, AccordionContent, AccordionItem, AccordionTrigger,
+} from "@/components/ui/accordion";
 import { Button } from "@/components/ui/button";
 import { Skeleton } from "@/components/ui/skeleton";
 import { CaseHeader } from "@/components/cases/case-header";
 import { CaseEditDialog } from "@/components/cases/case-edit-dialog";
-import { WorkflowTimeline } from "@/components/cases/workflow-timeline";
-import { StatusGuide } from "@/components/cases/status-guide";
-import { StatusChanger } from "@/components/cases/status-changer";
 import { DocumentsPanel } from "@/components/cases/documents-panel";
 import { DraftsPanel } from "@/components/cases/drafts-panel";
+import { CaseFilesBrowser } from "@/components/cases/case-files-browser";
 import { DecisionBlocksPanel } from "@/components/cases/decision-blocks-panel";
 import { LegalArgumentsPanel } from "@/components/cases/legal-arguments-panel";
+import { PositionsPanel } from "@/components/cases/positions-panel";
+import { HearingChangesPanel } from "@/components/cases/hearing-changes-panel";
+import { CitationVerificationPanel } from "@/components/compose/citation-verification-panel";
 import { AgentActivityFeed } from "@/components/cases/agent-activity-feed";
 import { AgentActivityPreview } from "@/components/cases/agent-activity-preview";
-import { AgentStatusWidget } from "@/components/cases/agent-status-widget";
 import { UploadSheet } from "@/components/documents/upload-sheet";
-import { expectedOutcomes } from "@/lib/schemas/case";
 import { useCase, useStartWorkflow } from "@/lib/api/cases";
 import { toast } from "sonner";
-import { Play, Loader2 } from "lucide-react";
-
-const EXPECTED_OUTCOME_LABELS: Record<string, string> = Object.fromEntries(
-  expectedOutcomes.map((o) => [o.value, o.label]),
-);
+import {
+  Play, Loader2, LayoutGrid, Scale, BadgeCheck, FileText, MessageSquare, Users,
+  type LucideIcon,
+} from "lucide-react";
 
 /*
  * Next 16 breaking change: route params are now a Promise.
@@ -39,12 +40,10 @@ export default function CaseDetailPage({
   params: Promise<{ caseNumber: string }>;
 }) {
   const { caseNumber } = use(params);
+  const [tab, setTab] = useState("overview");
   const { data, isPending, error, refetch } = useCase(caseNumber);
   const startWorkflow = useStartWorkflow(caseNumber);
   const canStartWorkflow = data?.status === "new" || data?.status === "documents_ready";
-  const expectedOutcomeLabel = data?.expected_outcome
-    ? EXPECTED_OUTCOME_LABELS[data.expected_outcome] ?? data.expected_outcome
-    : null;
 
   // Only take over the whole page when there is NO data to show. A transient
   // 5xx on the 5s background refetch must not blow away an already-loaded page.
@@ -71,23 +70,32 @@ export default function CaseDetailPage({
     );
   }
 
+  // Workflow order (X17): intake → arguments+positions → verify citations →
+  // write decision → drafts/final → agents (monitoring).
+  const tabDefs: [string, string, LucideIcon][] = [
+    ["overview", "סקירה", LayoutGrid],
+    ["arguments", "טיעונים ועמדות", Scale],
+    ["verify", "אימות פסיקה", BadgeCheck],
+    ["decision", "ההחלטה", FileText],
+    ["drafts", "טיוטות והערות", MessageSquare],
+    ["agents", "סוכנים", Users],
+  ];
+
+  // V2 "segmented" tab strip (X17): each tab a pill, active = raised white card.
+  // Spread full-width via justify-between (chair request) — wide gaps with few
+  // tabs, tightening as more are added; gap-2 is the no-overlap floor.
   const tabsList = (
     <TabsList
       variant="line"
-      className="gap-6 h-auto p-0 rounded-none -mb-px"
+      className="!h-auto w-full gap-2 p-0 justify-between"
     >
-      {[
-        ["overview", "סקירה"],
-        ["arguments", "טיעונים"],
-        ["decision", "ההחלטה"],
-        ["drafts", "טיוטות והערות"],
-        ["agents", "סוכנים"],
-      ].map(([value, label]) => (
+      {tabDefs.map(([value, label, Icon]) => (
         <TabsTrigger
           key={value}
           value={value}
-          className="flex-none rounded-none px-0 pb-3.5 pt-0 text-[0.92rem] font-medium text-ink-muted data-active:text-navy data-active:font-semibold data-active:after:bg-gold data-active:after:bottom-0"
+          className="flex-none gap-1.5 rounded-lg border border-transparent px-3.5 py-1.5 text-[0.9rem] font-medium text-ink-soft after:hidden hover:bg-parchment/60 data-active:bg-surface data-active:text-navy data-active:font-semibold data-active:border-rule data-active:shadow-sm [&_svg]:text-ink-light data-active:[&_svg]:text-gold-deep"
         >
+          <Icon className="w-4 h-4" />
           {label}
         </TabsTrigger>
       ))}
@@ -98,9 +106,6 @@ export default function CaseDetailPage({
     <>
       {data && <CaseEditDialog data={data} />}
       <UploadSheet caseNumber={caseNumber} />
-      <Button asChild className="bg-gold text-white hover:bg-gold-deep border-transparent">
-        <Link href={`/cases/${caseNumber}/compose`}>פתח עורך החלטה</Link>
-      </Button>
       {canStartWorkflow && (
         <Button
           className="bg-gold-deep hover:bg-gold-deep/90 text-parchment"
@@ -126,7 +131,7 @@ export default function CaseDetailPage({
 
   return (
     <AppShell>
-      <Tabs defaultValue="overview" dir="rtl">
+      <Tabs value={tab} onValueChange={setTab} dir="rtl">
         {/* parchment band — header (title/chips/parties/actions) + tab strip */}
         {isPending ? (
           <div className="-mx-10 -mt-10 mb-2 bg-parchment border-b border-rule px-10 pt-6 pb-4 space-y-3">
@@ -135,23 +140,98 @@ export default function CaseDetailPage({
             <Skeleton className="h-6 w-96" />
           </div>
         ) : (
-          <CaseHeader data={data} actions={bandActions} tabs={tabsList} />
+          <CaseHeader caseNumber={caseNumber} data={data} actions={bandActions} tabs={tabsList} />
         )}
 
-        {/* two-column wrap — main tab content (1fr) + rail (340px) */}
-        <div className="grid gap-6 lg:grid-cols-[1fr_340px] items-start mt-6">
-          <div className="min-w-0">
-            <TabsContent value="overview" className="mt-0 space-y-5">
+        {/* full-width tab content — status now lives in the always-visible band
+            StatusRail (banner A, X17), so every tab keeps the status context. */}
+        <div className="min-w-0 mt-6">
+          <TabsContent value="overview" className="mt-0">
+            {/* documents (wider) + agent-activity preview (mockup 18j) */}
+            <div className="grid gap-6 lg:grid-cols-[1.4fr_1fr] items-start">
               <DocumentsPanel data={data} />
-
               <AgentActivityPreview caseNumber={caseNumber} />
-              {/* decision-editor CTA moved to the band actions (visible on all tabs) */}
+            </div>
           </TabsContent>
 
-            <TabsContent value="arguments" className="mt-0">
+          {/* טיעונים ועמדות — chair positions (editing) over the analyst's
+              issues, with the aggregated by-party arguments collapsible below
+              (merged from the deleted /compose editor, mockup 18f). */}
+          <TabsContent value="arguments" className="mt-0 space-y-4">
+            {/* עמדות וטענות — collapsed by default (chair preference #226),
+                mirroring the two accordions below. */}
+            <Accordion type="single" collapsible>
+              <AccordionItem
+                value="positions"
+                className="overflow-hidden rounded-xl border border-rule bg-surface shadow-sm"
+              >
+                <AccordionTrigger className="px-6 py-4 hover:no-underline">
+                  <span className="flex flex-1 flex-col items-start">
+                    <span className="text-navy text-[0.95rem] font-bold">
+                      עמדות וטענות
+                    </span>
+                    <span className="text-ink-muted text-[0.78rem]">
+                      סוגיות-המחלוקת ועמדות הצדדים · עורך עמדת-היו״ר מזין את בלוק י׳
+                    </span>
+                  </span>
+                </AccordionTrigger>
+                <AccordionContent className="px-6 pb-5 pt-0">
+                  <PositionsPanel caseNumber={caseNumber} />
+                </AccordionContent>
+              </AccordionItem>
+            </Accordion>
+            <Accordion type="single" collapsible>
+              <AccordionItem
+                value="byparty"
+                className="overflow-hidden rounded-xl border border-rule bg-surface shadow-sm"
+              >
+                <AccordionTrigger className="px-6 py-4 hover:no-underline">
+                  <span className="flex flex-1 flex-col items-start">
+                    <span className="text-navy text-[0.95rem] font-bold">
+                      טיעונים מאוגדים לפי צד
+                    </span>
+                    <span className="text-ink-muted text-[0.78rem]">
+                      תצוגה מסונתזת של הטענות הגולמיות, מקובצת לפי צד וקדימות
+                    </span>
+                  </span>
+                </AccordionTrigger>
+                <AccordionContent className="px-6 pb-5 pt-0">
+                  <LegalArgumentsPanel caseNumber={caseNumber} />
+                </AccordionContent>
+              </AccordionItem>
+            </Accordion>
+            {/* מה קרה בדיון — comparative analysis of the hearing protocol vs.
+                the written pleadings (#226). Sits below the aggregated arguments
+                because it speaks to those same arguments in their oral gloss.
+                Collapsed by default, mirroring the "by-party" accordion above. */}
+            <Accordion type="single" collapsible>
+              <AccordionItem
+                value="hearing"
+                className="overflow-hidden rounded-xl border border-rule bg-surface shadow-sm"
+              >
+                <AccordionTrigger className="px-6 py-4 hover:no-underline">
+                  <span className="flex flex-1 flex-col items-start">
+                    <span className="text-navy text-[0.95rem] font-bold">
+                      מה קרה בדיון
+                    </span>
+                    <span className="text-ink-muted text-[0.78rem]">
+                      השוואת הטענות בכתב מול הדיון בעל-פה — מה התחזק ומה נטען לראשונה
+                    </span>
+                  </span>
+                </AccordionTrigger>
+                <AccordionContent className="px-6 pb-5 pt-0">
+                  <HearingChangesPanel caseNumber={caseNumber} />
+                </AccordionContent>
+              </AccordionItem>
+            </Accordion>
+          </TabsContent>
+
+          {/* אימות פסיקה — per-argument supporting-precedent verify gate (#154),
+              relocated from the deleted /compose editor to a top-level tab. */}
+          <TabsContent value="verify" className="mt-0">
             <Card className="bg-surface border-rule shadow-sm">
               <CardContent className="px-6 py-5">
-                  <LegalArgumentsPanel caseNumber={caseNumber} />
+                <CitationVerificationPanel caseNumber={caseNumber} />
               </CardContent>
             </Card>
           </TabsContent>
@@ -164,12 +244,17 @@ export default function CaseDetailPage({
             </Card>
           </TabsContent>
 
-            <TabsContent value="drafts" className="mt-0">
+          <TabsContent value="drafts" className="mt-0 space-y-4">
             <Card className="bg-surface border-rule shadow-sm">
               <CardContent className="px-6 py-5">
                 <DraftsPanel caseNumber={caseNumber} status={data?.status} />
               </CardContent>
             </Card>
+            <Card className="bg-surface border-rule shadow-sm">
+              <CardContent className="px-6 py-5">
+                <CaseFilesBrowser caseNumber={caseNumber} />
+              </CardContent>
+            </Card>
           </TabsContent>
 
           <TabsContent value="agents" className="mt-0">
@@ -180,32 +265,6 @@ export default function CaseDetailPage({
             </Card>
           </TabsContent>
         </div>
-
-          {/* rail — status timeline + status controls (mockup .rail) */}
-          <div className="space-y-5">
-            <Card className="bg-surface border-rule shadow-sm overflow-hidden p-0 h-fit">
-              <div className="px-5 py-3.5 border-b border-rule-soft bg-parchment text-[0.92rem] font-semibold text-navy">
-                סטטוס התיק
-              </div>
-              <CardContent className="px-5 py-4 space-y-4">
-                <AgentStatusWidget caseNumber={caseNumber} />
-                <WorkflowTimeline status={data?.status} />
-                {expectedOutcomeLabel && (
-                  <div className="border-t border-rule-soft pt-3">
-                    <dl className="flex justify-between items-center text-sm">
-                      <dt className="text-ink-muted">תוצאה צפויה</dt>
-                      <dd className="rounded-full bg-warn-bg text-warn text-[0.75rem] font-semibold px-3 py-0.5">
-                        {expectedOutcomeLabel}
-                      </dd>
-                    </dl>
-                  </div>
-                )}
-                <StatusChanger caseNumber={caseNumber} currentStatus={data?.status} />
-                <StatusGuide />
-              </CardContent>
-            </Card>
-          </div>
-        </div>
       </Tabs>
     </AppShell>
   );

web-ui/src/app/feedback/page.tsx

@@ -16,6 +16,7 @@ import {
   type ChairFeedback,
   type FeedbackCategory,
 } from "@/lib/api/feedback";
+import { formatDateLong } from "@/lib/format-date";
 
 /**
  * מרכז הערות יו"ר — הדף המרכזי לטיפול בכל הערות דפנה שנרשמו על טיוטות
@@ -33,18 +34,7 @@ const CAT_CHIP: Record<FeedbackCategory, string> = {
   other: "bg-rule-soft text-ink-soft",
 };
 
-function formatDate(iso?: string | null): string {
-  if (!iso) return "";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL", {
-      day: "numeric",
-      month: "long",
-      year: "numeric",
-    });
-  } catch {
-    return "";
-  }
-}
+const formatDate = (iso?: string | null) => formatDateLong(iso);
 
 function FeedbackCard({ fb }: { fb: ChairFeedback }) {
   const resolve = useResolveFeedback();

web-ui/src/app/operations/page.tsx

@@ -29,13 +29,17 @@ import {
   useRunLog,
   useCancelRun,
   useResetAgentSession,
+  useAgentHealth,
+  useRecentEscalations,
   type OpsService,
   type OperationsSnapshot,
   type PipelineStats,
   type AgentRun,
   type SubscriptionUsage,
   type UsageWindow,
+  type AgentHealthState,
 } from "@/lib/api/operations";
+import { formatTime, israelParts } from "@/lib/format-date";
 
 function mb(bytes: number): string {
   return `${Math.round((bytes || 0) / 1024 / 1024)}MB`;
@@ -46,8 +50,7 @@ function mb(bytes: number): string {
 // and resumes on its own when a window resets.
 function usageResetLabel(iso: string | null): string {
   if (!iso) return "—";
-  const d = new Date(iso);
-  return `איפוס ${d.toLocaleTimeString("he-IL", { hour: "2-digit", minute: "2-digit" })}`;
+  return `איפוס ${formatTime(iso)}`;
 }
 
 function UsageMeter({ label, w }: { label: string; w?: UsageWindow | null }) {
@@ -195,8 +198,9 @@ function nextSaturday18(): Date {
 }
 const HE_DAYS = ["א׳", "ב׳", "ג׳", "ד׳", "ה׳", "ו׳", "ש׳"];
 function fmtDeadline(iso: string): string {
-  const d = new Date(iso);
-  return `${HE_DAYS[d.getDay()]} ${pad2(d.getDate())}/${pad2(d.getMonth() + 1)} ${pad2(d.getHours())}:${pad2(d.getMinutes())}`;
+  const p = israelParts(iso);
+  if (!p) return "—";
+  return `${HE_DAYS[p.weekday]} ${pad2(p.day)}/${pad2(p.month)} ${pad2(p.hour)}:${pad2(p.minute)}`;
 }
 
 function BurstControl({ s }: { s: OpsService }) {
@@ -801,6 +805,154 @@ function LiveAgentsPanel() {
   );
 }
 
+// ── Agent health taxonomy (#222) + recent escalations (#218) ───────────────
+const HEALTH_LABEL: Record<AgentHealthState, string> = {
+  zombie: "זומבי",
+  stalled: "תקוע",
+  working: "עובד",
+  idle: "רגוע",
+};
+
+function AgentHealthPanel() {
+  const { data, isLoading } = useAgentHealth();
+  const { data: esc } = useRecentEscalations();
+
+  const counts = data?.counts;
+  // Only zombies/stalled are "problems" — the backend already sorts worst-first.
+  // An escalated issue is chair-owned, so it leaves this list; the escalations
+  // card below is the handled-loop history.
+  const problems = (data?.items ?? []).filter(
+    (i) => i.health === "zombie" || i.health === "stalled",
+  );
+  const healthyWorking = counts?.working ?? 0;
+  const healthyIdle = counts?.idle ?? 0;
+
+  return (
+    <div className="space-y-4">
+      <Card className="bg-surface border-rule shadow-sm">
+        <CardContent className="px-6 py-5">
+          {isLoading || !data ? (
+            <Skeleton className="h-24 w-full" />
+          ) : (
+            <>
+              <div className="flex flex-wrap gap-2 mb-4">
+                <StatTile
+                  label="זומבי"
+                  value={counts?.zombie ?? 0}
+                  tone="red"
+                  title="לולאת-recovery — משויך+פעיל-כביכול אך אין ריצה חיה. מוסלם אוטומטית."
+                />
+                <StatTile
+                  label="תקוע"
+                  value={counts?.stalled ?? 0}
+                  tone="amber"
+                  title="ננער חוזר בלי התקדמות"
+                />
+                <StatTile label="עובד" value={counts?.working ?? 0} tone="green" />
+                <StatTile label="רגוע" value={counts?.idle ?? 0} tone="muted" />
+              </div>
+
+              {problems.length === 0 ? (
+                <p className="text-sm text-emerald-600">
+                  אין סוכן תקוע כרגע — כל הסוכנים תקינים.
+                </p>
+              ) : (
+                <div className="grid gap-2">
+                  {problems.map((p) => (
+                    <div
+                      key={p.issue_id}
+                      className={
+                        p.health === "zombie"
+                          ? "flex items-center justify-between gap-3 rounded-md border border-destructive/30 bg-destructive/5 px-3 py-2"
+                          : "flex items-center justify-between gap-3 rounded-md border border-gold/40 bg-gold-wash px-3 py-2"
+                      }
+                    >
+                      <div className="flex items-center gap-2 flex-wrap min-w-0">
+                        <Badge
+                          variant={p.health === "zombie" ? "destructive" : "outline"}
+                          className={
+                            p.health === "stalled"
+                              ? "border-warn/50 text-warn font-normal"
+                              : "font-normal"
+                          }
+                        >
+                          {HEALTH_LABEL[p.health]}
+                        </Badge>
+                        <span dir="ltr" className="font-mono text-[0.82rem] text-navy font-semibold">
+                          {p.identifier}
+                        </span>
+                        <span className="text-[0.82rem] text-navy font-semibold">
+                          {p.agent_name}
+                        </span>
+                        <span className="text-[0.7rem] text-ink-muted">
+                          {p.recovery_wakeups > 0
+                            ? `${p.recovery_wakeups} יקיצות-שחזור`
+                            : `${p.wakeups} יקיצות`}
+                        </span>
+                      </div>
+                      <span className="text-[0.7rem] text-ink-muted shrink-0">
+                        {p.health === "zombie" ? "יוסלם אוטומטית" : "מנוטר"}
+                      </span>
+                    </div>
+                  ))}
+                </div>
+              )}
+
+              <p className="mt-3 text-[0.78rem] text-ink-muted">
+                <b className="text-navy">{healthyWorking + healthyIdle} סוכנים תקינים</b> —{" "}
+                {healthyWorking} עובדים, {healthyIdle} רגועים. מוצגות רק בעיות.
+              </p>
+            </>
+          )}
+        </CardContent>
+      </Card>
+
+      {/* handled-loop history — the watchdog's escalations, gold-wash like a gate */}
+      <Card className="bg-gold-wash border-gold/40 shadow-sm">
+        <CardContent className="px-5 py-4">
+          <h3 className="text-navy text-sm font-semibold mb-0.5">הסלמות אוטומטיות אחרונות</h3>
+          <p className="text-ink-muted text-[0.72rem] mb-3">
+            כל זומבי מתמשך (לולאת-recovery ≥2 יקיצות-שחזור) משויך אליך אוטומטית עם הערת-מערכת
+            — בלי להעיר סוכן מחדש.
+          </p>
+          {!esc?.items?.length ? (
+            <p className="text-sm text-ink-muted">אין הסלמות אחרונות.</p>
+          ) : (
+            <div className="grid gap-1">
+              {esc.items.map((e) => (
+                <div
+                  key={`${e.issue_id}-${e.created_at ?? ""}`}
+                  className="flex items-start gap-2.5 border-t border-gold/15 first:border-0 pt-2 first:pt-0"
+                >
+                  <Badge
+                    variant={e.severity === "medium" ? "outline" : "destructive"}
+                    className={
+                      e.severity === "medium"
+                        ? "border-gold/50 text-gold-deep font-normal shrink-0"
+                        : "font-normal shrink-0"
+                    }
+                  >
+                    {e.severity}
+                  </Badge>
+                  <div className="min-w-0 flex-1">
+                    <span dir="ltr" className="font-mono text-[0.8rem] text-navy font-semibold">
+                      {e.identifier}
+                    </span>
+                    <div className="text-[0.78rem] text-ink-soft">{e.reason}</div>
+                  </div>
+                  <span className="text-[0.68rem] text-ink-muted shrink-0">
+                    {e.created_at ? ago(Date.parse(e.created_at)) : ""}
+                  </span>
+                </div>
+              ))}
+            </div>
+          )}
+        </CardContent>
+      </Card>
+    </div>
+  );
+}
+
 export default function OperationsPage() {
   const { data, isLoading, error } = useOperations();
 
@@ -839,6 +991,9 @@ export default function OperationsPage() {
             <SectionHeader>סוכנים פעילים</SectionHeader>
             <LiveAgentsPanel />
 
+            <SectionHeader>בריאות-הסוכנים</SectionHeader>
+            <AgentHealthPanel />
+
             <SectionHeader>מתאמי-סוכנים</SectionHeader>
             <AgentAdaptersPanel />
 

web-ui/src/app/scripts/page.tsx

@@ -26,6 +26,7 @@ import {
   useRunScript,
   type ScriptRunResult,
 } from "@/lib/api/scripts";
+import { formatDateLong } from "@/lib/format-date";
 
 /*
  * /scripts — catalog of everything under scripts/, rendered as the
@@ -354,11 +355,7 @@ export default function ScriptsPage() {
 
   const lastModified =
     data?.last_modified != null
-      ? new Date(data.last_modified * 1000).toLocaleDateString("he-IL", {
-          year: "numeric",
-          month: "long",
-          day: "numeric",
-        })
+      ? formatDateLong(data.last_modified * 1000)
       : null;
 
   const giteaBase = data?.gitea_url ?? null;

web-ui/src/app/settings/_components/agents-tab.tsx

@@ -19,6 +19,7 @@ import {
   type DriftEntry,
   type PaperclipAgent,
 } from "@/lib/api/paperclip-agents";
+import { formatDateShort } from "@/lib/format-date";
 
 const ROLE_LABEL: Record<string, string> = {
   ceo: "CEO",
@@ -211,7 +212,7 @@ function PairCard({ pair }: { pair: AgentPair }) {
           </Button>
           {pair.master?.updated_at && (
             <span className="text-[0.7rem] text-ink-light">
-              עודכן: {new Date(pair.master.updated_at).toLocaleDateString("he-IL")}
+              עודכן: {formatDateShort(pair.master.updated_at)}
             </span>
           )}
         </div>

web-ui/src/app/skills/page.tsx

@@ -13,6 +13,7 @@ import {
   TableRow,
 } from "@/components/ui/table";
 import { useSkills, type Skill } from "@/lib/api/skills";
+import { formatIsoDate } from "@/lib/format-date";
 
 function formatChars(skill: Skill): string {
   // Mockup column = "גודל (תווים)" — the DB markdown char count, grouped.
@@ -21,13 +22,8 @@ function formatChars(skill: Skill): string {
 }
 
 function formatUpdated(iso: string | null): string {
-  if (!iso) return "—";
-  try {
-    // ISO-style date to match the mockup (2026-06-09), tabular.
-    return new Date(iso).toISOString().slice(0, 10);
-  } catch {
-    return "—";
-  }
+  // ISO-style date to match the mockup (2026-06-09), tabular — Israel-time day.
+  return formatIsoDate(iso);
 }
 
 /**

web-ui/src/components/cases/agent-activity-feed.tsx

@@ -1,6 +1,6 @@
 "use client";
 
-import { useRef, useState, useEffect, useMemo } from "react";
+import { useState, useMemo } from "react";
 import { Button } from "@/components/ui/button";
 import { Textarea } from "@/components/ui/textarea";
 import { Badge } from "@/components/ui/badge";
@@ -9,6 +9,8 @@ import {
   useAgentActivity,
   useSendComment,
   useSubmitInteraction,
+  useDismissInteraction,
+  useSetIssueStatus,
 } from "@/lib/api/agents";
 import type {
   Interaction,
@@ -16,7 +18,10 @@ import type {
   InteractionQuestion,
   InteractionTask,
   PaperclipComment,
+  PaperclipAgentStatus,
+  PaperclipIssue,
 } from "@/lib/api/agents";
+import { formatRelative } from "@/lib/format-date";
 import { toast } from "sonner";
 import {
   Bot,
@@ -28,6 +33,11 @@ import {
   CheckCircle2,
   XCircle,
   HelpCircle,
+  ChevronDown,
+  Ban,
+  Target,
+  Plus,
+  AlertTriangle,
 } from "lucide-react";
 
 /* ── Role → color mapping ────────────────────────────────────── */
@@ -80,18 +90,83 @@ function issueStatusLabel(status: string) {
   return ISSUE_STATUS_LABELS[status] ?? status;
 }
 
+/* ── Agent roster (mockup 18i) — who's doing what right now ────── */
+
+const AGENT_STATUS_LABEL: Record<string, string> = {
+  active: "פעיל",
+  running: "פעיל",
+  idle: "ממתין",
+  error: "שגיאה",
+};
+
+function agentIsActive(s: string) {
+  return s === "active" || s === "running";
+}
+
+function agentStatusDot(s: string) {
+  return agentIsActive(s)
+    ? "bg-success shadow-[0_0_6px_rgba(74,124,89,0.6)]"
+    : s === "error"
+      ? "bg-danger"
+      : "bg-rule";
+}
+
+function agentStatusTone(s: string) {
+  return agentIsActive(s)
+    ? "bg-success-bg text-success"
+    : s === "error"
+      ? "bg-danger-bg text-danger"
+      : "bg-rule-soft text-ink-muted";
+}
+
+function AgentRoster({ agents }: { agents: PaperclipAgentStatus[] }) {
+  if (!agents.length) return null;
+
+  const activeCount = agents.filter((a) => agentIsActive(a.status)).length;
+
+  return (
+    <div className="mb-3 rounded-lg border border-rule bg-parchment/40 p-3">
+      <div className="flex items-center gap-2 mb-2.5">
+        <Bot className="w-4 h-4 text-gold-deep shrink-0" />
+        <h3 className="text-[0.82rem] font-bold text-navy">סוכני התיק</h3>
+        <span className="text-[0.72rem] text-ink-muted">
+          <b className="text-success font-bold">{activeCount} פעילים</b> מתוך {agents.length}
+        </span>
+      </div>
+      {/* Compact roster — all agents on one row (mockup 18i v2): status · name ·
+          role only. The per-agent "current activity" line was removed: it
+          matched comments by role-fallback, so agents sharing a role showed the
+          same text. */}
+      <div className="grid gap-1.5 grid-cols-3 sm:grid-cols-5 lg:grid-cols-9">
+        {agents.map((a) => {
+          const active = agentIsActive(a.status);
+          return (
+            <div
+              key={a.id}
+              className={`rounded-lg border p-2 min-w-0 ${active ? "border-success/40 bg-success-bg/40" : "border-rule bg-surface"}`}
+            >
+              <div className="flex items-center gap-1.5 min-w-0">
+                <span className={`w-1.5 h-1.5 rounded-full shrink-0 ${agentStatusDot(a.status)}`} />
+                <span className="text-[0.72rem] font-bold text-navy truncate">{a.name}</span>
+              </div>
+              <div className="text-[0.6rem] text-ink-muted mt-0.5 truncate">{roleLabel(a.role)}</div>
+              <span
+                className={`inline-block mt-1 text-[0.55rem] font-bold rounded-full px-1.5 py-0.5 whitespace-nowrap ${agentStatusTone(a.status)}`}
+              >
+                {AGENT_STATUS_LABEL[a.status] ?? a.status}
+              </span>
+            </div>
+          );
+        })}
+      </div>
+    </div>
+  );
+}
+
 /* ── Time formatting ─────────────────────────────────────────── */
 
 function timeAgo(iso: string | null): string {
-  if (!iso) return "";
-  const diff = Date.now() - new Date(iso).getTime();
-  const mins = Math.floor(diff / 60_000);
-  if (mins < 1) return "עכשיו";
-  if (mins < 60) return `לפני ${mins} דק׳`;
-  const hours = Math.floor(mins / 60);
-  if (hours < 24) return `לפני ${hours} שע׳`;
-  const days = Math.floor(hours / 24);
-  return `לפני ${days} ימים`;
+  return formatRelative(iso, { units: "short", fallback: "" });
 }
 
 /* ── Issue identifier → find matching identifier ─────────────── */
@@ -513,15 +588,21 @@ function InteractionCard({
   interaction,
   caseNumber,
   issueMap,
+  defaultOpen = true,
 }: {
   interaction: Interaction;
   caseNumber: string;
   issueMap: Map<string, string>;
+  defaultOpen?: boolean;
 }) {
   const submit = useSubmitInteraction(caseNumber);
+  const dismiss = useDismissInteraction(caseNumber);
   const identifier = issueMap.get(interaction.issue_id) ?? "";
   const isPending = interaction.status === "pending";
   const summary = summaryAnswer(interaction);
+  // Pending interactions collapse into an accordion (mockup 18i v2); resolved
+  // ones always render their one-line summary.
+  const [open, setOpen] = useState(defaultOpen);
 
   const send = (action: "respond" | "accept" | "reject", payload: InteractionPayload | Record<string, unknown>) => {
     submit.mutate(
@@ -538,6 +619,16 @@ function InteractionCard({
     );
   };
 
+  const handleDismiss = () => {
+    dismiss.mutate(
+      { issue_id: interaction.issue_id, interaction_id: interaction.id },
+      {
+        onSuccess: () => toast.success("הבקשה בוטלה"),
+        onError: () => toast.error("שגיאה בביטול הבקשה"),
+      },
+    );
+  };
+
   return (
     <div
       className={`group relative flex gap-3 py-3 px-2 rounded-lg border transition-colors ${
@@ -560,15 +651,30 @@ function InteractionCard({
 
       <div className="flex-1 min-w-0">
         <div className="flex items-center gap-2 mb-2 flex-wrap">
+          {isPending ? (
+            <button
+              type="button"
+              onClick={() => setOpen((o) => !o)}
+              aria-expanded={open}
+              className="flex items-center gap-2 flex-wrap text-start min-w-0"
+            >
+              <ChevronDown
+                className={`w-4 h-4 text-ink-faint shrink-0 transition-transform ${open ? "rotate-180" : ""}`}
+              />
               <span className="text-sm font-semibold text-navy">
                 {interaction.title || "שאלה לסוכן"}
               </span>
-          {isPending ? (
               <span className="inline-flex items-center gap-1 text-[11px] px-2 py-0.5 rounded-full border border-amber-300 bg-amber-100 text-amber-800">
                 ממתין לתשובה
               </span>
+            </button>
           ) : (
+            <>
+              <span className="text-sm font-semibold text-navy">
+                {interaction.title || "שאלה לסוכן"}
+              </span>
               <ResolvedBadge status={interaction.status} />
+            </>
           )}
           {identifier && (
             <Badge variant="outline" className="text-[10px] font-mono">
@@ -579,13 +685,31 @@ function InteractionCard({
             <Clock className="w-3 h-3" />
             {timeAgo(interaction.resolved_at ?? interaction.created_at)}
           </span>
+          {isPending && (
+            <Button
+              size="sm"
+              variant="ghost"
+              onClick={handleDismiss}
+              disabled={dismiss.isPending}
+              className="h-7 px-2 text-ink-muted hover:text-danger"
+              title="בטל בקשה כפולה/מיושנת — ללא העברה לסוכן"
+            >
+              {dismiss.isPending ? (
+                <Loader2 className="w-3.5 h-3.5 animate-spin" />
+              ) : (
+                <Ban className="w-3.5 h-3.5 me-1" />
+              )}
+              התעלם
+            </Button>
+          )}
         </div>
 
-        {interaction.summary && (
+        {interaction.summary && (!isPending || open) && (
           <div className="text-xs text-ink-faint mb-2">{interaction.summary}</div>
         )}
 
         {isPending ? (
+          open ? (
             interaction.kind === "ask_user_questions" ? (
               <AskUserQuestionsForm
                 interaction={interaction}
@@ -613,6 +737,7 @@ function InteractionCard({
                 pending={submit.isPending}
               />
             ) : null
+          ) : null
         ) : summary ? (
           <div className="text-sm">
             <Markdown content={summary} />
@@ -623,6 +748,480 @@ function InteractionCard({
   );
 }
 
+/* ── Issue group (timeline grouped by issue, accordion) ──────── */
+
+type FeedItem =
+  | { kind: "comment"; at: number; comment: PaperclipComment }
+  | { kind: "interaction"; at: number; interaction: Interaction };
+
+/** Drop the "[ערר 1043-02-26] " prefix Paperclip puts on every issue title. */
+function shortIssueTitle(title: string): string {
+  return title.replace(/^\s*\[[^\]]*\]\s*/, "").trim() || title;
+}
+
+function IssueGroup({
+  issue,
+  items,
+  caseNumber,
+  issueMap,
+  defaultOpen,
+}: {
+  issue: PaperclipIssue;
+  items: FeedItem[];
+  caseNumber: string;
+  issueMap: Map<string, string>;
+  defaultOpen: boolean;
+}) {
+  const [open, setOpen] = useState(defaultOpen);
+  const setStatus = useSetIssueStatus(caseNumber);
+  const closed = issue.status === "done" || issue.status === "cancelled";
+  // Show the parent run this issue hangs under (e.g. a "הרצה חדשה" nested under
+  // the live CEO root) so the hierarchy is visible, not flattened.
+  const parentIdentifier = issue.parent_id
+    ? issueMap.get(issue.parent_id)
+    : undefined;
+
+  const close = (status: "done" | "cancelled") =>
+    setStatus.mutate(
+      { issue_id: issue.id, status },
+      {
+        onSuccess: () =>
+          toast.success(
+            status === "done" ? "המשימה סומנה כהושלמה" : "המשימה בוטלה",
+          ),
+        onError: () => toast.error("שגיאה בעדכון סטטוס המשימה"),
+      },
+    );
+
+  return (
+    <div className="border-b-2 border-rule last:border-b-0">
+      <div className="w-full flex items-center gap-2 px-2 py-2.5 hover:bg-sand-soft/50">
+        <button
+          type="button"
+          onClick={() => setOpen((o) => !o)}
+          aria-expanded={open}
+          className="flex-1 flex items-center gap-2 text-start min-w-0"
+        >
+          <Badge
+            variant={closed ? "secondary" : "default"}
+            className="text-[10px] font-mono shrink-0"
+          >
+            {issue.identifier}
+          </Badge>
+          <span className="text-[0.78rem] font-semibold text-navy truncate">
+            {shortIssueTitle(issue.title)}
+          </span>
+          {parentIdentifier && (
+            <span
+              className="inline-flex items-center gap-0.5 text-[0.6rem] font-bold text-gold-deep bg-gold/10 border border-rule rounded-full px-1.5 py-0.5 shrink-0 whitespace-nowrap"
+              title={`תת-משימה של ${parentIdentifier}`}
+            >
+              ↳ {parentIdentifier}
+            </span>
+          )}
+          <span className="text-[0.7rem] text-ink-muted whitespace-nowrap">
+            {issueStatusLabel(issue.status)} · {items.length}
+          </span>
+        </button>
+        {!closed && (
+          <div className="flex items-center gap-1 shrink-0">
+            <Button
+              size="sm"
+              variant="outline"
+              onClick={() => close("done")}
+              disabled={setStatus.isPending}
+              className="h-7 px-2 text-[0.7rem] text-success border-success/40 hover:bg-success-bg"
+              title="סמן את המשימה כהושלמה"
+            >
+              {setStatus.isPending ? (
+                <Loader2 className="w-3 h-3 animate-spin" />
+              ) : (
+                <CheckCircle2 className="w-3.5 h-3.5 me-1" />
+              )}
+              הושלם
+            </Button>
+            <Button
+              size="sm"
+              variant="ghost"
+              onClick={() => close("cancelled")}
+              disabled={setStatus.isPending}
+              className="h-7 px-2 text-[0.7rem] text-ink-muted hover:text-danger"
+              title="בטל את המשימה"
+            >
+              בטל
+            </Button>
+          </div>
+        )}
+        <button
+          type="button"
+          onClick={() => setOpen((o) => !o)}
+          aria-label={open ? "כווץ" : "הרחב"}
+          className="shrink-0"
+        >
+          <ChevronDown
+            className={`w-4 h-4 text-ink-faint transition-transform ${open ? "rotate-180" : ""}`}
+          />
+        </button>
+      </div>
+      {open && (
+        <div className="pb-1">
+          {items.map((item) =>
+            item.kind === "comment" ? (
+              <CommentCard
+                key={`c-${item.comment.id}`}
+                comment={item.comment}
+                issueMap={issueMap}
+              />
+            ) : (
+              <InteractionCard
+                key={`i-${item.interaction.id}`}
+                interaction={item.interaction}
+                caseNumber={caseNumber}
+                issueMap={issueMap}
+              />
+            ),
+          )}
+        </div>
+      )}
+    </div>
+  );
+}
+
+/* ── Compose: directive box + target selector (top of tab) ────── */
+
+const isOpenStatus = (s: string) => s !== "done" && s !== "cancelled";
+
+/** Client mirror of the server's pick_default_comment_target
+ * (web/paperclip_client.py): prefer the newest OPEN top-level issue (the live
+ * CEO main issue), then any open issue, then any top-level, else newest.
+ * `issues` arrives oldest→newest, so the last match is the newest. */
+function pickDefaultTarget(issues: PaperclipIssue[]): PaperclipIssue | null {
+  if (!issues.length) return null;
+  const isTop = (i: PaperclipIssue) => i.parent_id == null;
+  const preds: ((i: PaperclipIssue) => boolean)[] = [
+    (i) => isOpenStatus(i.status) && isTop(i),
+    (i) => isOpenStatus(i.status),
+    isTop,
+  ];
+  for (const p of preds) {
+    const m = issues.filter(p);
+    if (m.length) return m[m.length - 1];
+  }
+  return issues[issues.length - 1];
+}
+
+const STATUS_TONE: Record<string, string> = {
+  in_progress: "bg-emerald-100 text-emerald-700",
+  in_review: "bg-amber-100 text-amber-700",
+  todo: "bg-emerald-100 text-emerald-700",
+  backlog: "bg-emerald-100 text-emerald-700",
+  blocked: "bg-red-100 text-red-700",
+  done: "bg-gray-100 text-gray-500",
+  cancelled: "bg-red-50 text-red-600",
+};
+function statusTone(s: string) {
+  return STATUS_TONE[s] ?? "bg-gray-100 text-gray-600";
+}
+
+/** Where a chair instruction is routed. `auto` lets the server pick the live
+ * CEO main issue; `issue` targets an explicit one; `new_run` opens a fresh
+ * CEO-owned run (bypasses the human-owned-issue cancellation). */
+type Target =
+  | { kind: "auto" }
+  | { kind: "issue"; id: string }
+  | { kind: "new_run" };
+
+function TargetSelector({
+  issues,
+  target,
+  onChange,
+}: {
+  issues: PaperclipIssue[];
+  target: Target;
+  onChange: (t: Target) => void;
+}) {
+  const [open, setOpen] = useState(false);
+  const defaultIssue = useMemo(() => pickDefaultTarget(issues), [issues]);
+  const activeIssues = issues.filter((i) => isOpenStatus(i.status));
+  const closedIssues = issues.filter((i) => !isOpenStatus(i.status));
+  // id → identifier, to label a run with the parent it hangs under (↳ CMP-189).
+  const idToIdentifier = useMemo(
+    () => new Map(issues.map((i) => [i.id, i.identifier])),
+    [issues],
+  );
+  const parentTag = (i: PaperclipIssue) =>
+    i.parent_id ? idToIdentifier.get(i.parent_id) : undefined;
+
+  // The issue the pill currently represents (explicit pick, or the auto default).
+  const effectiveIssue =
+    target.kind === "issue"
+      ? issues.find((i) => i.id === target.id) ?? null
+      : target.kind === "auto"
+        ? defaultIssue
+        : null;
+
+  return (
+    <div className="relative min-w-0">
+      <button
+        type="button"
+        onClick={() => setOpen((o) => !o)}
+        aria-expanded={open}
+        className="flex items-center gap-2 bg-sand-soft/60 border border-rule rounded-lg px-2.5 py-1.5 hover:border-gold/60 min-w-0 max-w-full"
+      >
+        <Target className="w-3.5 h-3.5 text-gold-deep shrink-0" />
+        {target.kind === "new_run" ? (
+          <span className="text-xs font-semibold text-navy">פתח הרצה חדשה</span>
+        ) : effectiveIssue ? (
+          <>
+            <Badge variant="outline" className="text-[10px] font-mono shrink-0">
+              {effectiveIssue.identifier}
+            </Badge>
+            <span className="text-xs font-medium text-navy truncate">
+              {shortIssueTitle(effectiveIssue.title)}
+            </span>
+            <span
+              className={`text-[10px] font-semibold rounded-full px-2 py-0.5 shrink-0 ${statusTone(effectiveIssue.status)}`}
+            >
+              {issueStatusLabel(effectiveIssue.status)}
+            </span>
+          </>
+        ) : (
+          <span className="text-xs text-ink-muted">בחר יעד</span>
+        )}
+        <ChevronDown className="w-3.5 h-3.5 text-ink-faint shrink-0 ms-1" />
+      </button>
+
+      {open && (
+        <>
+          <div
+            className="fixed inset-0 z-10"
+            onClick={() => setOpen(false)}
+            aria-hidden
+          />
+          <div className="absolute z-20 top-full mt-1 start-0 w-[360px] max-w-[88vw] max-h-[62vh] overflow-y-auto overscroll-contain bg-white border border-rule rounded-xl shadow-lg p-1.5">
+            {/* "פתח הרצה חדשה" is the FIRST, most-prominent option so it's always
+                visible without scrolling to the bottom of a long task list. */}
+            <button
+              type="button"
+              onClick={() => {
+                onChange({ kind: "new_run" });
+                setOpen(false);
+              }}
+              className="w-full flex items-center gap-2.5 px-2 py-2 rounded-lg text-start bg-gold/10 border border-rule hover:border-gold/60"
+            >
+              <span className="w-6 h-6 rounded-md bg-navy text-white flex items-center justify-center shrink-0">
+                <Plus className="w-4 h-4" />
+              </span>
+              <span className="min-w-0">
+                <span className="block text-xs font-semibold text-navy">
+                  פתח הרצה חדשה
+                </span>
+                <span className="block text-[10px] text-ink-muted">
+                  ה-CEO ייצור משימה חדשה ויטפל בהוראה מאפס
+                </span>
+              </span>
+            </button>
+            <div className="h-px bg-rule-soft my-1.5 mx-1" />
+
+            <div className="px-2 py-1 text-[10px] font-bold text-ink-faint uppercase tracking-wide">
+              משימות פעילות
+            </div>
+            {activeIssues.map((i) => {
+              const selected =
+                (target.kind === "issue" && target.id === i.id) ||
+                (target.kind === "auto" && defaultIssue?.id === i.id);
+              const parent = parentTag(i);
+              return (
+                <button
+                  key={i.id}
+                  type="button"
+                  onClick={() => {
+                    onChange({ kind: "issue", id: i.id });
+                    setOpen(false);
+                  }}
+                  className={`w-full flex items-center gap-2 px-2 py-2 rounded-lg text-start hover:bg-sand-soft ${parent ? "ps-6" : ""} ${selected ? "bg-gold/10" : ""}`}
+                >
+                  <span className="w-1.5 h-1.5 rounded-full bg-emerald-500 shrink-0" />
+                  <Badge variant="outline" className="text-[10px] font-mono shrink-0">
+                    {i.identifier}
+                  </Badge>
+                  {parent && (
+                    <span
+                      className="text-[9px] font-bold text-gold-deep bg-gold/10 border border-rule rounded-full px-1.5 py-0.5 shrink-0 whitespace-nowrap"
+                      title={`תת-משימה של ${parent}`}
+                    >
+                      ↳ {parent}
+                    </span>
+                  )}
+                  <span className="text-xs text-ink-soft truncate flex-1">
+                    {shortIssueTitle(i.title)}
+                  </span>
+                  <span
+                    className={`text-[10px] font-semibold rounded-full px-2 py-0.5 shrink-0 ${statusTone(i.status)}`}
+                  >
+                    {issueStatusLabel(i.status)}
+                  </span>
+                  {selected && (
+                    <CheckCircle2 className="w-3.5 h-3.5 text-gold-deep shrink-0" />
+                  )}
+                </button>
+              );
+            })}
+
+            {closedIssues.length > 0 && (
+              <>
+                <div className="h-px bg-rule-soft my-1.5 mx-1" />
+                <div className="px-2 py-1 text-[10px] font-bold text-ink-faint uppercase tracking-wide">
+                  משימות סגורות · שליחה אליהן לא תעיר סוכן
+                </div>
+                {closedIssues.map((i) => {
+                  const parent = parentTag(i);
+                  return (
+                    <div
+                      key={i.id}
+                      className={`w-full flex items-center gap-2 px-2 py-2 rounded-lg opacity-60 cursor-not-allowed ${parent ? "ps-6" : ""}`}
+                      title="משימה סגורה — שליחה אליה לא תעיר סוכן"
+                    >
+                      <span className="w-1.5 h-1.5 rounded-full bg-ink-faint shrink-0" />
+                      <Badge variant="outline" className="text-[10px] font-mono shrink-0">
+                        {i.identifier}
+                      </Badge>
+                      {parent && (
+                        <span
+                          className="text-[9px] font-bold text-gold-deep bg-gold/10 border border-rule rounded-full px-1.5 py-0.5 shrink-0 whitespace-nowrap"
+                          title={`תת-משימה של ${parent}`}
+                        >
+                          ↳ {parent}
+                        </span>
+                      )}
+                      <span className="text-xs text-ink-faint truncate flex-1">
+                        {shortIssueTitle(i.title)}
+                      </span>
+                      <span
+                        className={`text-[10px] font-semibold rounded-full px-2 py-0.5 shrink-0 ${statusTone(i.status)}`}
+                      >
+                        {issueStatusLabel(i.status)}
+                      </span>
+                    </div>
+                  );
+                })}
+              </>
+            )}
+          </div>
+        </>
+      )}
+    </div>
+  );
+}
+
+function Composer({
+  caseNumber,
+  issues,
+}: {
+  caseNumber: string;
+  issues: PaperclipIssue[];
+}) {
+  const sendComment = useSendComment(caseNumber);
+  const [body, setBody] = useState("");
+  const [target, setTarget] = useState<Target>({ kind: "auto" });
+
+  const selectedIssue =
+    target.kind === "issue" ? issues.find((i) => i.id === target.id) ?? null : null;
+  const closedTarget = !!selectedIssue && !isOpenStatus(selectedIssue.status);
+
+  const handleSend = () => {
+    if (!body.trim() || closedTarget) return;
+    const vars =
+      target.kind === "new_run"
+        ? { body: body.trim(), new_run: true }
+        : target.kind === "issue"
+          ? { body: body.trim(), issue_id: target.id }
+          : { body: body.trim() };
+    sendComment.mutate(vars, {
+      onSuccess: (res) => {
+        setBody("");
+        setTarget({ kind: "auto" });
+        toast.success(
+          res.new_run
+            ? `נפתחה הרצה חדשה — ${res.issue_identifier}`
+            : `נשלח ל-${res.issue_identifier}`,
+        );
+      },
+      onError: () => toast.error("שגיאה בשליחת ההודעה"),
+    });
+  };
+
+  return (
+    <div className="bg-white border border-rule rounded-xl shadow-sm p-3.5">
+      <div className="flex items-center gap-2 mb-2.5">
+        <MessageSquare className="w-4 h-4 text-gold-deep" />
+        <h3 className="text-sm font-bold text-navy">הוראה לסוכנים</h3>
+        <span className="text-[11px] text-ink-faint ms-auto">
+          Ctrl+Enter לשליחה
+        </span>
+      </div>
+
+      <div className="flex items-center gap-2 mb-2 flex-wrap">
+        <span className="text-[11px] font-semibold text-ink-muted shrink-0">
+          שליחה אל:
+        </span>
+        <TargetSelector issues={issues} target={target} onChange={setTarget} />
+      </div>
+
+      {closedTarget && (
+        <div className="flex items-start gap-2 bg-red-50 border border-red-200 rounded-lg px-3 py-2 mb-2 text-[11.5px] text-red-700 leading-snug">
+          <AlertTriangle className="w-4 h-4 shrink-0 mt-0.5" />
+          <div>
+            <b className="font-bold">יעד סגור.</b> שליחה למשימה סגורה לא תעיר סוכן —
+            ההוראה לא תטופל. בחר משימה פעילה או{" "}
+            <button
+              type="button"
+              className="font-bold underline"
+              onClick={() => setTarget({ kind: "new_run" })}
+            >
+              פתח הרצה חדשה
+            </button>
+            .
+          </div>
+        </div>
+      )}
+
+      <Textarea
+        value={body}
+        onChange={(e) => setBody(e.target.value)}
+        placeholder="כתוב הוראה לסוכנים..."
+        aria-label="הוראה לסוכנים"
+        className="min-h-[60px] resize-none text-sm"
+        dir="rtl"
+        onKeyDown={(e) => {
+          if (e.key === "Enter" && (e.ctrlKey || e.metaKey)) {
+            e.preventDefault();
+            handleSend();
+          }
+        }}
+      />
+      <div className="flex items-center mt-2">
+        <span className="text-[11px] text-ink-faint">
+          ההודעה תנותב דרך סוכן ה-CEO · היעד ניתן לשינוי למעלה
+        </span>
+        <Button
+          size="sm"
+          className="ms-auto"
+          onClick={handleSend}
+          disabled={!body.trim() || closedTarget || sendComment.isPending}
+        >
+          {sendComment.isPending ? (
+            <Loader2 className="w-4 h-4 animate-spin" />
+          ) : (
+            <Send className="w-4 h-4 me-1" />
+          )}
+          שלח
+        </Button>
+      </div>
+    </div>
+  );
+}
+
 /* ── Main Feed ───────────────────────────────────────────────── */
 
 export function AgentActivityFeed({
@@ -631,9 +1230,6 @@ export function AgentActivityFeed({
   caseNumber: string;
 }) {
   const { data, isLoading, error } = useAgentActivity(caseNumber);
-  const sendComment = useSendComment(caseNumber);
-  const [body, setBody] = useState("");
-  const endRef = useRef<HTMLDivElement>(null);
 
   // Build issue_id → identifier map (memoized — the feed refetches every 10s,
   // and a fresh Map each render would defeat the child cards' memoization).
@@ -643,27 +1239,6 @@ export function AgentActivityFeed({
     return m;
   }, [data?.issues]);
 
-  // Auto-scroll on new comments or interactions
-  const commentCount = data?.comments?.length ?? 0;
-  const interactionCount = data?.interactions?.length ?? 0;
-  useEffect(() => {
-    endRef.current?.scrollIntoView({ behavior: "smooth" });
-  }, [commentCount, interactionCount]);
-
-  const handleSend = () => {
-    if (!body.trim()) return;
-    sendComment.mutate(
-      { body: body.trim() },
-      {
-        onSuccess: (res) => {
-          setBody("");
-          toast.success(`נשלח ל-${res.issue_identifier}`);
-        },
-        onError: () => toast.error("שגיאה בשליחת ההודעה"),
-      },
-    );
-  };
-
   // ── Empty / loading states ──
 
   if (isLoading) {
@@ -696,24 +1271,46 @@ export function AgentActivityFeed({
 
   const comments = data.comments ?? [];
   const interactions = data.interactions ?? [];
+  // Pending interactions are surfaced at the top (awaiting-you, mockup 18i) and
+  // NOT repeated in the timeline below.
+  const pending = interactions.filter((i) => i.status === "pending");
 
-  // Unified, time-sorted feed: comments + interactions interleaved.
-  type FeedItem =
-    | { kind: "comment"; at: number; comment: PaperclipComment }
-    | { kind: "interaction"; at: number; interaction: Interaction };
-
-  const feed: FeedItem[] = [
-    ...comments.map<FeedItem>((c) => ({
+  // Group the timeline by issue, each an accordion (mockup 18i v2). Comments +
+  // resolved interactions; pending ones are surfaced above and not repeated.
+  const itemsByIssue = new Map<string, FeedItem[]>();
+  const pushItem = (issueId: string, item: FeedItem) => {
+    const arr = itemsByIssue.get(issueId);
+    if (arr) arr.push(item);
+    else itemsByIssue.set(issueId, [item]);
+  };
+  for (const c of comments) {
+    pushItem(c.issue_id, {
       kind: "comment",
       at: c.created_at ? new Date(c.created_at).getTime() : 0,
       comment: c,
-    })),
-    ...interactions.map<FeedItem>((i) => ({
+    });
+  }
+  for (const i of interactions) {
+    if (i.status === "pending") continue;
+    pushItem(i.issue_id, {
       kind: "interaction",
       at: i.created_at ? new Date(i.created_at).getTime() : 0,
       interaction: i,
-    })),
-  ].sort((a, b) => a.at - b.at);
+    });
+  }
+  for (const arr of itemsByIssue.values()) arr.sort((a, b) => a.at - b.at);
+
+  // Order groups by most-recent activity so live work floats to the top.
+  const issueGroups = data.issues
+    .map((iss) => {
+      const items = itemsByIssue.get(iss.id) ?? [];
+      const maxAt = items.reduce((m, it) => Math.max(m, it.at), 0);
+      return { iss, items, maxAt };
+    })
+    .filter((g) => g.items.length > 0)
+    .sort((a, b) => b.maxAt - a.maxAt);
+
+  const totalFeedItems = issueGroups.reduce((n, g) => n + g.items.length, 0);
 
   // An issue is "active" if it's not done/cancelled. When everything is closed
   // we should NOT show the "agents are working, waiting for report" spinner.
@@ -722,23 +1319,51 @@ export function AgentActivityFeed({
   );
 
   return (
-    <div className="flex flex-col h-full">
-      {/* Issue summary bar */}
-      <div className="flex items-center gap-2 px-2 py-2 border-b border-rule mb-2 flex-wrap">
-        {data.issues.map((iss) => (
-          <Badge
-            key={iss.id}
-            variant={iss.status === "done" ? "secondary" : "default"}
-            className="text-[11px] font-mono"
-          >
-            {iss.identifier} — {issueStatusLabel(iss.status)}
-          </Badge>
+    <div className="flex flex-col h-full gap-3">
+      {/* directive box — moved to the TOP, above the activity (most available,
+          no scrolling). Carries the target selector. */}
+      <Composer caseNumber={caseNumber} issues={data.issues} />
+
+      {/* agent roster — who's who, compact one-row (mockup 18i v2) */}
+      <AgentRoster agents={data.agents ?? []} />
+
+      {/* pending interactions surfaced to the top — each an accordion, all
+          COLLAPSED by default (chair opens to look). Excluded from the timeline. */}
+      {pending.length > 0 && (
+        <div className="rounded-lg border border-amber-300 bg-amber-50 p-3 space-y-2">
+          <div className="flex items-center gap-2 text-[0.82rem] font-bold text-amber-800">
+            <Clock className="w-4 h-4" />
+            ממתין לתשובתך
+            <span className="text-[0.7rem] font-semibold bg-amber-100 rounded-full px-2 py-0.5">
+              {pending.length}
+            </span>
+          </div>
+          {pending.map((i) => (
+            <InteractionCard
+              key={`pending-${i.id}`}
+              interaction={i}
+              caseNumber={caseNumber}
+              issueMap={issueMap}
+              defaultOpen={false}
+            />
           ))}
         </div>
+      )}
 
-      {/* Comments + interactions stream */}
-      <div className="flex-1 overflow-y-auto max-h-[500px] space-y-1 px-1">
-        {feed.length === 0 ? (
+      {/* Timeline header */}
+      <div className="flex items-center gap-2 px-2 py-2 border-b border-rule">
+        <span className="text-[0.82rem] font-bold text-navy">פעילות לפי משימה</span>
+        <span className="text-[0.68rem] font-semibold text-gold-deep bg-gold/10 rounded-full px-2 py-0.5">
+          החדש למעלה
+        </span>
+        <span className="text-[0.72rem] text-ink-muted ms-auto">
+          {data.issues.length} משימות · {totalFeedItems} הודעות
+        </span>
+      </div>
+
+      {/* Per-issue accordions — all COLLAPSED by default, newest group first */}
+      <div className="flex-1 overflow-y-auto max-h-[500px]">
+        {totalFeedItems === 0 ? (
           hasActiveIssue ? (
             <div className="text-center py-8 text-ink-faint text-sm">
               <Loader2 className="w-5 h-5 animate-spin mx-auto mb-2" />
@@ -751,59 +1376,17 @@ export function AgentActivityFeed({
             </div>
           )
         ) : (
-          feed.map((item) =>
-            item.kind === "comment" ? (
-              <CommentCard
-                key={`c-${item.comment.id}`}
-                comment={item.comment}
-                issueMap={issueMap}
-              />
-            ) : (
-              <InteractionCard
-                key={`i-${item.interaction.id}`}
-                interaction={item.interaction}
+          issueGroups.map(({ iss, items }) => (
+            <IssueGroup
+              key={iss.id}
+              issue={iss}
+              items={items}
               caseNumber={caseNumber}
               issueMap={issueMap}
+              defaultOpen={false}
             />
-            ),
-          )
+          ))
         )}
-        <div ref={endRef} />
-      </div>
-
-      {/* Comment input */}
-      <div className="border-t border-rule pt-3 mt-3 space-y-2">
-        <Textarea
-          value={body}
-          onChange={(e) => setBody(e.target.value)}
-          placeholder="כתוב הוראה לסוכנים..."
-          aria-label="הוראה לסוכנים"
-          className="min-h-[60px] resize-none text-sm"
-          dir="rtl"
-          onKeyDown={(e) => {
-            if (e.key === "Enter" && (e.ctrlKey || e.metaKey)) {
-              e.preventDefault();
-              handleSend();
-            }
-          }}
-        />
-        <div className="flex items-center justify-between">
-          <span className="text-[11px] text-ink-faint">
-            ההודעה תנותב דרך סוכן ה-CEO &middot; Ctrl+Enter לשליחה
-          </span>
-          <Button
-            size="sm"
-            onClick={handleSend}
-            disabled={!body.trim() || sendComment.isPending}
-          >
-            {sendComment.isPending ? (
-              <Loader2 className="w-4 h-4 animate-spin" />
-            ) : (
-              <Send className="w-4 h-4 me-1" />
-            )}
-            שלח
-          </Button>
-        </div>
       </div>
     </div>
   );

web-ui/src/components/cases/agent-activity-preview.tsx

@@ -4,6 +4,7 @@ import Link from "next/link";
 import { Loader2, CheckCircle2, Clock } from "lucide-react";
 import { useAgentActivity } from "@/lib/api/agents";
 import type { PaperclipIssue } from "@/lib/api/agents";
+import { formatRelative } from "@/lib/format-date";
 
 const STATUS_DOT: Record<string, string> = {
   in_progress: "bg-gold",
@@ -22,13 +23,7 @@ const STATUS_LABEL: Record<string, string> = {
 };
 
 function timeAgo(iso: string | null): string {
-  if (!iso) return "—";
-  const diffMs = Date.now() - new Date(iso).getTime();
-  const m = Math.floor(diffMs / 60_000);
-  if (m < 60) return `לפני ${m}ד'`;
-  const h = Math.floor(m / 60);
-  if (h < 24) return `לפני ${h}ש'`;
-  return `לפני ${Math.floor(h / 24)} ימים`;
+  return formatRelative(iso, { units: "tight", fallback: "—" });
 }
 
 function IssueRow({ issue }: { issue: PaperclipIssue }) {

web-ui/src/components/cases/case-files-browser.tsx

@@ -0,0 +1,304 @@
+"use client";
+
+import { useMemo, useState } from "react";
+import {
+  useCaseFiles,
+  useCaseFileText,
+  caseFileUrl,
+  type CaseFileFolder,
+} from "@/lib/api/case-files";
+import { Markdown } from "@/components/ui/markdown";
+import { Badge } from "@/components/ui/badge";
+import { formatDateTime } from "@/lib/format-date";
+import {
+  Folder,
+  FileText,
+  File as FileIcon,
+  Download,
+  ChevronDown,
+  Loader2,
+  FolderOpen,
+} from "lucide-react";
+
+/* ── Helpers ─────────────────────────────────────────────────── */
+
+const FOLDER_LABELS: Record<string, string> = {
+  research: "מחקר וטיוטות",
+  originals: "מסמכי מקור",
+  extracted: "טקסט מחולץ (OCR)",
+  proofread: "לאחר הגהה",
+  drafts: "טיוטות DOCX",
+  backup: "גיבויים",
+};
+
+// Extensions the inline viewer can render (everything else → download only).
+const TEXT_EXT = new Set(["md", "txt", "json", "csv", "log", "markdown"]);
+
+function fileExt(filename: string): string {
+  return filename.includes(".") ? filename.split(".").pop()!.toLowerCase() : "";
+}
+
+function folderLabel(f: CaseFileFolder): string {
+  return FOLDER_LABELS[f.name] ?? f.name;
+}
+
+function fmtSize(bytes: number): string {
+  if (bytes < 1024) return `${bytes} B`;
+  if (bytes < 1024 * 1024) return `${(bytes / 1024).toFixed(1)} KB`;
+  return `${(bytes / 1024 / 1024).toFixed(1)} MB`;
+}
+
+const EXT_TONE: Record<string, string> = {
+  md: "bg-teal-100 text-teal-700",
+  markdown: "bg-teal-100 text-teal-700",
+  txt: "bg-blue-100 text-blue-700",
+  json: "bg-blue-100 text-blue-700",
+  pdf: "bg-red-100 text-red-700",
+  docx: "bg-indigo-100 text-indigo-700",
+  doc: "bg-indigo-100 text-indigo-700",
+};
+function extTone(ext: string): string {
+  return EXT_TONE[ext] ?? "bg-gray-100 text-gray-600";
+}
+
+type Selected = { folderKey: string; filename: string };
+
+/* ── Component ───────────────────────────────────────────────── */
+
+export function CaseFilesBrowser({ caseNumber }: { caseNumber: string }) {
+  const { data, isLoading, error } = useCaseFiles(caseNumber);
+  const folders = useMemo(() => data?.folders ?? [], [data?.folders]);
+
+  const totalFiles = folders.reduce((n, f) => n + f.files.length, 0);
+
+  // Default selection derived (not via an effect): a text file in research,
+  // else any text file. The chair's explicit pick overrides it.
+  const defaultSelected = useMemo<Selected | null>(() => {
+    const pick = (
+      pred: (f: CaseFileFolder, name: string) => boolean,
+    ): Selected | null => {
+      for (const f of folders)
+        for (const file of f.files)
+          if (pred(f, file.filename))
+            return { folderKey: f.key, filename: file.filename };
+      return null;
+    };
+    return (
+      pick((f, name) => f.name === "research" && TEXT_EXT.has(fileExt(name))) ??
+      pick((_f, name) => TEXT_EXT.has(fileExt(name)))
+    );
+  }, [folders]);
+
+  const [userSelected, setUserSelected] = useState<Selected | null>(null);
+  const selected = userSelected ?? defaultSelected;
+
+  // A folder is open if explicitly toggled; otherwise the folder of the current
+  // selection (or the first non-empty folder) is open by default.
+  const defaultOpenKey =
+    selected?.folderKey ?? folders.find((f) => f.files.length)?.key ?? null;
+  const [openOverrides, setOpenOverrides] = useState<Map<string, boolean>>(
+    new Map(),
+  );
+  const isOpen = (key: string) =>
+    openOverrides.has(key) ? openOverrides.get(key)! : key === defaultOpenKey;
+
+  const selExt = selected ? fileExt(selected.filename) : "";
+  const selIsText = TEXT_EXT.has(selExt);
+  const {
+    data: fileText,
+    isLoading: textLoading,
+    error: textError,
+  } = useCaseFileText(
+    caseNumber,
+    selIsText ? selected?.folderKey : undefined,
+    selIsText ? selected?.filename : undefined,
+  );
+
+  const toggleFolder = (key: string) => {
+    const next = !isOpen(key);
+    setOpenOverrides((prev) => new Map(prev).set(key, next));
+  };
+
+  if (isLoading) {
+    return (
+      <div className="flex items-center justify-center py-10 text-ink-faint">
+        <Loader2 className="w-5 h-5 animate-spin me-2" />
+        <span>טוען קבצים...</span>
+      </div>
+    );
+  }
+  if (error) {
+    return (
+      <div className="text-center py-10 text-red-500 text-sm">
+        שגיאה בטעינת קבצי-התיק
+      </div>
+    );
+  }
+
+  return (
+    <div className="border border-rule rounded-xl overflow-hidden">
+      {/* header */}
+      <div className="flex items-center gap-2 px-4 py-3 border-b border-rule-soft bg-parchment">
+        <FolderOpen className="w-4 h-4 text-gold-deep" />
+        <h3 className="text-sm font-bold text-navy">קבצי-התיק</h3>
+        <span className="text-[11px] text-ink-muted ms-auto">
+          כל תתי-התיקיות · {totalFiles} קבצים
+        </span>
+      </div>
+
+      <div className="grid md:grid-cols-[320px_1fr]">
+        {/* file tree */}
+        <div className="border-b md:border-b-0 md:border-e border-rule-soft max-h-[560px] overflow-y-auto">
+          {folders.map((f) => {
+            const open = isOpen(f.key);
+            return (
+              <div key={f.key} className="border-b border-rule-soft last:border-b-0">
+                <button
+                  type="button"
+                  onClick={() => toggleFolder(f.key)}
+                  aria-expanded={open}
+                  className="w-full flex items-center gap-2 px-3.5 py-2.5 hover:bg-sand-soft/60 text-start"
+                >
+                  <Folder className="w-4 h-4 text-gold-deep shrink-0" />
+                  <span className="text-[0.78rem] font-bold text-navy">
+                    {folderLabel(f)}
+                  </span>
+                  <span className="text-[9px] font-mono text-ink-faint hidden sm:inline">
+                    {f.path}
+                  </span>
+                  <span
+                    className={`text-[10px] font-semibold rounded-full px-2 ms-auto ${f.files.length ? "bg-gray-100 text-ink-muted" : "text-ink-light"}`}
+                  >
+                    {f.files.length || "ריק"}
+                  </span>
+                  <ChevronDown
+                    className={`w-4 h-4 text-ink-faint shrink-0 transition-transform ${open ? "rotate-180" : ""}`}
+                  />
+                </button>
+
+                {open && f.files.length > 0 && (
+                  <div className="pb-1 bg-parchment/50">
+                    {f.files.map((file) => {
+                      const ext = fileExt(file.filename);
+                      const isSel =
+                        selected?.folderKey === f.key &&
+                        selected?.filename === file.filename;
+                      const viewable = TEXT_EXT.has(ext);
+                      return (
+                        <div
+                          key={file.filename}
+                          className={`flex items-center gap-2.5 py-1.5 pe-3.5 ps-5 cursor-pointer border-s-2 ${isSel ? "bg-gold-wash border-gold" : "border-transparent hover:bg-sand-soft/60"}`}
+                          onClick={() =>
+                            viewable
+                              ? setUserSelected({
+                                  folderKey: f.key,
+                                  filename: file.filename,
+                                })
+                              : undefined
+                          }
+                        >
+                          <span
+                            className={`w-6 h-6 rounded-md flex items-center justify-center shrink-0 ${extTone(ext)}`}
+                          >
+                            {viewable ? (
+                              <FileText className="w-3.5 h-3.5" />
+                            ) : (
+                              <FileIcon className="w-3.5 h-3.5" />
+                            )}
+                          </span>
+                          <div className="min-w-0 flex-1">
+                            <div className="text-[11.5px] font-medium text-ink-soft truncate">
+                              {file.filename}
+                            </div>
+                            <div className="text-[9.5px] text-ink-light">
+                              {fmtSize(file.size)} ·{" "}
+                              {formatDateTime(file.modified_at * 1000)}
+                            </div>
+                          </div>
+                          <a
+                            href={caseFileUrl(caseNumber, f.key, file.filename)}
+                            download={file.filename}
+                            onClick={(e) => e.stopPropagation()}
+                            className="w-7 h-7 rounded-md border border-rule bg-surface flex items-center justify-center text-ink-muted hover:text-navy shrink-0"
+                            title="הורדה"
+                          >
+                            <Download className="w-3.5 h-3.5" />
+                          </a>
+                        </div>
+                      );
+                    })}
+                  </div>
+                )}
+              </div>
+            );
+          })}
+          {folders.length === 0 && (
+            <div className="text-center py-10 text-ink-faint text-sm">
+              אין קבצים בתיק.
+            </div>
+          )}
+        </div>
+
+        {/* viewer */}
+        <div className="min-w-0 flex flex-col">
+          {!selected ? (
+            <div className="flex-1 flex flex-col items-center justify-center py-16 text-ink-faint text-sm">
+              <FileText className="w-8 h-8 mb-2 opacity-40" />
+              בחר קובץ טקסט מהרשימה לצפייה
+            </div>
+          ) : (
+            <>
+              <div className="flex items-center gap-2.5 px-4 py-3 border-b border-rule-soft">
+                <span
+                  className={`w-7 h-7 rounded-md flex items-center justify-center shrink-0 ${extTone(selExt)}`}
+                >
+                  <FileText className="w-4 h-4" />
+                </span>
+                <div className="min-w-0">
+                  <div className="text-[12.5px] font-bold text-navy truncate">
+                    {selected.filename}
+                  </div>
+                </div>
+                <a
+                  href={caseFileUrl(
+                    caseNumber,
+                    selected.folderKey,
+                    selected.filename,
+                  )}
+                  download={selected.filename}
+                  className="ms-auto inline-flex items-center gap-1.5 text-[11.5px] font-semibold text-navy bg-parchment border border-rule rounded-md px-3 py-1.5 hover:bg-sand-soft"
+                >
+                  <Download className="w-3.5 h-3.5" />
+                  הורדה
+                </a>
+              </div>
+              <div className="px-5 py-4 overflow-y-auto max-h-[520px] bg-[#fdfcf9]">
+                {!selIsText ? (
+                  <div className="text-center py-12 text-sm text-ink-muted">
+                    קובץ מסוג <Badge variant="outline">{selExt || "?"}</Badge> —
+                    לא ניתן לצפייה מוטבעת. הורד כדי לפתוח.
+                  </div>
+                ) : textLoading ? (
+                  <div className="flex items-center justify-center py-12 text-ink-faint">
+                    <Loader2 className="w-5 h-5 animate-spin me-2" />
+                    טוען תוכן...
+                  </div>
+                ) : textError ? (
+                  <div className="text-center py-12 text-red-500 text-sm">
+                    שגיאה בטעינת תוכן הקובץ
+                  </div>
+                ) : selExt === "md" || selExt === "markdown" ? (
+                  <Markdown content={fileText ?? ""} />
+                ) : (
+                  <pre className="whitespace-pre-wrap text-xs text-ink-soft font-mono leading-relaxed">
+                    {fileText}
+                  </pre>
+                )}
+              </div>
+            </>
+          )}
+        </div>
+      </div>
+    </div>
+  );
+}

web-ui/src/components/cases/case-header.tsx

@@ -2,7 +2,7 @@ import type { ReactNode } from "react";
 import Link from "next/link";
 import { Badge } from "@/components/ui/badge";
 import { StatusBadge } from "@/components/cases/status-badge";
-import { SyncIndicator } from "@/components/cases/sync-indicator";
+import { StatusRail } from "@/components/cases/status-rail";
 import { CaseArchiveAction } from "@/components/cases/case-archive-action";
 import { CreateRepoButton } from "@/components/cases/create-repo-button";
 import {
@@ -10,30 +10,14 @@ import {
   APPEAL_SUBTYPE_LABELS,
   isBlamSubtype,
 } from "@/lib/practice-area";
+import {
+  Popover,
+  PopoverContent,
+  PopoverTrigger,
+} from "@/components/ui/popover";
+import { Users } from "lucide-react";
 import type { CaseDetail } from "@/lib/api/cases";
 
-function formatDate(iso?: string | null) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL", {
-      day: "2-digit",
-      month: "2-digit",
-      year: "numeric",
-    });
-  } catch {
-    return iso ?? "—";
-  }
-}
-
-function partiesLine(data?: CaseDetail): string | null {
-  const appellant = data?.appellants?.filter(Boolean) ?? [];
-  const respondent = data?.respondents?.filter(Boolean) ?? [];
-  const parts: string[] = [];
-  if (appellant.length) parts.push(`עוררת: ${appellant.join(", ")}`);
-  if (respondent.length) parts.push(`משיבה: ${respondent.join(", ")}`);
-  return parts.length ? parts.join(" · ") : null;
-}
-
 /**
  * Case header — parchment band (IA-redesign mockup 17): full-bleed band with
  * the case title + status/type chips inline, a parties line, the case actions
@@ -41,15 +25,19 @@ function partiesLine(data?: CaseDetail): string | null {
  * the tab strip inside the band, anchored to its bottom edge.
  */
 export function CaseHeader({
+  caseNumber,
   data,
   actions,
   tabs,
 }: {
+  caseNumber: string;
   data?: CaseDetail;
   actions?: ReactNode;
   tabs?: ReactNode;
 }) {
-  const parties = partiesLine(data);
+  const appellants = data?.appellants?.filter(Boolean) ?? [];
+  const respondents = data?.respondents?.filter(Boolean) ?? [];
+  const hasParties = appellants.length > 0 || respondents.length > 0;
   const isBlam =
     data?.proceeding_type === 'בל"מ' || isBlamSubtype(data?.appeal_subtype);
 
@@ -72,6 +60,12 @@ export function CaseHeader({
             <span className="tabular-nums">
               {data?.proceeding_type ?? "ערר"} {data?.case_number ?? "—"}
             </span>
+            {/* case subject on the same line, smaller font (chair request) */}
+            {data?.title && (
+              <span className="text-navy/75 text-lg font-semibold leading-snug">
+                {data.title}
+              </span>
+            )}
             {data?.status && <StatusBadge status={data.status} />}
             {data?.archived_at && (
               <Badge
@@ -103,27 +97,44 @@ export function CaseHeader({
             )}
           </h1>
 
-          {/* case title / subject under the heading */}
-          {data?.title && (
-            <p className="text-navy/90 text-base font-semibold mt-2 max-w-3xl leading-snug">
-              {data.title}
-            </p>
-          )}
-          {/* parties line (mockup .parties) — clamped to 2 lines so a long list
-              of appellants/respondents doesn't grow the band; full text on hover. */}
-          {parties ? (
-            <p className="text-ink-soft text-sm mt-1.5 max-w-3xl line-clamp-2" title={parties}>
-              {parties}
-            </p>
-          ) : data?.subject ? (
-            <p className="text-ink-soft text-sm mt-1.5 max-w-3xl leading-relaxed line-clamp-2"
-               title={data.subject}>
-              {data.subject}
-            </p>
-          ) : null}
-
-          {/* case actions — kept verbatim, moved into the band */}
+          {/* case actions — parties now live in a popover (chair request) so
+              they don't push the tab content down the page. */}
           <div className="flex items-center gap-2 flex-wrap mt-3">
+            {hasParties && (
+              <Popover>
+                <PopoverTrigger asChild>
+                  <button
+                    type="button"
+                    className="inline-flex items-center gap-1.5 rounded-md border border-rule bg-surface px-3 py-1.5 text-[0.82rem] font-semibold text-ink-soft hover:bg-parchment transition-colors"
+                  >
+                    <Users className="w-3.5 h-3.5" />
+                    צדדים
+                  </button>
+                </PopoverTrigger>
+                <PopoverContent align="start" className="w-72 p-3 space-y-2.5 text-sm">
+                  {appellants.length > 0 && (
+                    <div>
+                      <div className="text-[0.7rem] font-bold text-gold-deep uppercase tracking-wider mb-1">
+                        עוררים
+                      </div>
+                      <ul className="space-y-0.5 text-ink-soft">
+                        {appellants.map((a, i) => <li key={`a${i}`}>{a}</li>)}
+                      </ul>
+                    </div>
+                  )}
+                  {respondents.length > 0 && (
+                    <div>
+                      <div className="text-[0.7rem] font-bold text-gold-deep uppercase tracking-wider mb-1">
+                        משיבים
+                      </div>
+                      <ul className="space-y-0.5 text-ink-soft">
+                        {respondents.map((r, i) => <li key={`r${i}`}>{r}</li>)}
+                      </ul>
+                    </div>
+                  )}
+                </PopoverContent>
+              </Popover>
+            )}
             {data?.case_number && (
               <CaseArchiveAction
                 caseNumber={data.case_number}
@@ -135,25 +146,13 @@ export function CaseHeader({
           </div>
         </div>
 
-        {/* metadata strip — hearing date / updated / sync */}
-        <dl className="grid grid-cols-2 gap-x-6 gap-y-1 text-sm shrink-0">
-          <dt className="text-ink-muted text-[0.72rem] uppercase tracking-wider">
-            תאריך דיון
-          </dt>
-          <dd className="text-ink-soft tabular-nums">{formatDate(data?.hearing_date)}</dd>
-          <dt className="text-ink-muted text-[0.72rem] uppercase tracking-wider">
-            עודכן
-          </dt>
-          <dd className="text-ink-soft tabular-nums">{formatDate(data?.updated_at)}</dd>
-          <dt className="text-ink-muted text-[0.72rem] uppercase tracking-wider">
-            סנכרון
-          </dt>
-          <dd><SyncIndicator caseNumber={data?.case_number} /></dd>
-        </dl>
       </div>
 
-      {/* tab strip anchored to band bottom (mockup .tabs) */}
-      {tabs ? <div className="mt-5">{tabs}</div> : null}
+      {/* ★ integrated status rail — visible on every tab (banner A, X17) */}
+      <StatusRail caseNumber={caseNumber} data={data} />
+
+      {/* tab strip anchored to band bottom */}
+      {tabs ? <div className="mt-4">{tabs}</div> : null}
     </div>
   );
 }

web-ui/src/components/cases/cases-table.tsx

@@ -24,32 +24,16 @@ import {
 import { StatusBadge } from "@/components/cases/status-badge";
 import { isBlamSubtype } from "@/lib/practice-area";
 import type { Case } from "@/lib/api/cases";
+import { formatDate, israelMidnightMs, todayIsraelMidnightMs } from "@/lib/format-date";
 
-function formatDate(iso?: string) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL", {
-      day: "2-digit",
-      month: "2-digit",
-      year: "numeric",
-    });
-  } catch {
-    return iso;
-  }
-}
-
-/** Midnight today, in ms — the pivot between an upcoming and a past hearing. */
+/** Midnight today in Israel time, in ms — pivot between an upcoming and a past hearing. */
 function startOfToday() {
-  const d = new Date();
-  d.setHours(0, 0, 0, 0);
-  return d.getTime();
+  return todayIsraelMidnightMs();
 }
 
-/** Day-precision ms for a hearing date, or null when absent/unparseable. */
+/** Day-precision ms (Israel time) for a hearing date, or null when absent/unparseable. */
 function hearingMs(iso?: string | null): number | null {
-  if (!iso) return null;
-  const t = new Date(iso).setHours(0, 0, 0, 0);
-  return Number.isNaN(t) ? null : t;
+  return israelMidnightMs(iso);
 }
 
 type HearingClass = "soon" | "upcoming" | "past" | "none";

web-ui/src/components/cases/decision-blocks-panel.tsx

@@ -17,6 +17,7 @@ import {
   type BlockStatus,
 } from "@/lib/api/decision-blocks";
 import { BLOCK_LABELS } from "@/lib/api/feedback";
+import { formatTime } from "@/lib/format-date";
 import { AlertTriangle, Pencil, FileText } from "lucide-react";
 
 /* ── status badge styling ─────────────────────────────── */
@@ -233,10 +234,7 @@ function SaveIndicator({ state }: { state: SaveState }) {
     return <span className="text-[0.72rem] text-ink-muted">⏳ שומר…</span>;
   }
   if (state.kind === "saved") {
-    const time = state.at.toLocaleTimeString("he-IL", {
-      hour: "2-digit",
-      minute: "2-digit",
-    });
+    const time = formatTime(state.at);
     return <span className="text-[0.72rem] text-success">✓ נשמר {time}</span>;
   }
   return <span className="text-[0.72rem] text-danger">⚠ {state.message}</span>;

web-ui/src/components/cases/document-type-editor.tsx

@@ -4,6 +4,7 @@ import { useState } from "react";
 import { CheckCircle2, Loader2, Sparkles } from "lucide-react";
 import { Badge } from "@/components/ui/badge";
 import { Button } from "@/components/ui/button";
+import { Switch } from "@/components/ui/switch";
 import {
   Popover,
   PopoverContent,
@@ -20,6 +21,7 @@ import {
   APPRAISER_SIDE_LABELS,
   APPRAISER_SIDE_OPTIONS,
   DOC_TYPE_OPTIONS,
+  PROTOCOL_SCOPE_OPTIONS,
   appraiserSideLabel,
   doctypeLabel,
   doctypeTone,
@@ -29,14 +31,17 @@ import {
 import {
   useExtractAppraiserFacts,
   usePatchDocument,
+  type DocumentPatch,
   type ExtractAppraiserFactsResponse,
 } from "@/lib/api/documents";
 
 /*
  * Inline editor for a document's tags. Renders a colored Badge that opens a
- * Popover with two Selects:
+ * Popover with:
  *   1. doc_type (always shown)
  *   2. appraiser_side  (only when doc_type === "appraisal")
+ *   3. protocol_scope  (only when doc_type === "protocol") — ועדת הערר / מקומית
+ *   4. is_post_hearing toggle (always shown) — moves the doc to overview group 4
  *
  * After a successful save we swap the Popover body to a confirmation view
  * with a "חלץ עובדות שמאיות עכשיו" button — extraction is expensive so we
@@ -49,15 +54,25 @@ export function DocumentTypeEditor({
   docId,
   docType,
   appraiserSide,
+  isPostHearing,
+  protocolScope,
 }: {
   caseNumber: string;
   docId: string;
   docType: string;
   appraiserSide?: string;
+  isPostHearing?: boolean;
+  protocolScope?: string;
 }) {
+  // Effective stored scope: absent === "appeal" (ועדת הערר) by convention.
+  const storedScope = protocolScope === "lower" ? "lower" : "appeal";
+  const storedPost = isPostHearing === true;
+
   const [open, setOpen] = useState(false);
   const [draftType, setDraftType] = useState<string>(docType || "");
   const [draftSide, setDraftSide] = useState<string>(appraiserSide || "");
+  const [draftScope, setDraftScope] = useState<string>(storedScope);
+  const [draftPost, setDraftPost] = useState<boolean>(storedPost);
   const [saved, setSaved] = useState(false);
   const [extractResult, setExtractResult] =
     useState<ExtractAppraiserFactsResponse | null>(null);
@@ -67,6 +82,8 @@ export function DocumentTypeEditor({
   function reset() {
     setDraftType(docType || "");
     setDraftSide(appraiserSide || "");
+    setDraftScope(storedScope);
+    setDraftPost(storedPost);
     setSaved(false);
     setExtractResult(null);
     patch.reset();
@@ -74,14 +91,17 @@ export function DocumentTypeEditor({
   }
 
   const isAppraisal = draftType === "appraisal";
+  const isProtocol = draftType === "protocol";
   const sideMissing = isAppraisal && !draftSide;
   const dirty =
     draftType !== docType ||
-    (isAppraisal && draftSide !== (appraiserSide || ""));
+    (isAppraisal && draftSide !== (appraiserSide || "")) ||
+    (isProtocol && draftScope !== storedScope) ||
+    draftPost !== storedPost;
 
   async function handleSave() {
     if (sideMissing || !dirty) return;
-    const body: { doc_type?: string; appraiser_side?: string } = {};
+    const body: DocumentPatch = {};
     if (draftType !== docType) body.doc_type = draftType;
     if (isAppraisal && draftSide !== (appraiserSide || "")) {
       body.appraiser_side = draftSide;
@@ -90,6 +110,16 @@ export function DocumentTypeEditor({
     // clear it so it doesn't dangle confusingly in metadata.
     if (!isAppraisal && appraiserSide) body.appraiser_side = "";
 
+    // protocol_scope: store only the non-default "lower"; anything else clears
+    // back to the implicit ועדת-הערר default. Also clear when leaving protocol.
+    if (isProtocol && draftScope !== storedScope) {
+      body.protocol_scope = draftScope === "lower" ? "lower" : "";
+    } else if (!isProtocol && protocolScope) {
+      body.protocol_scope = "";
+    }
+
+    if (draftPost !== storedPost) body.is_post_hearing = draftPost;
+
     // Swallow the rejection — errors surface via `patch.isError`; an unhandled
     // rejection from this async click handler would otherwise leak.
     try {
@@ -190,6 +220,46 @@ export function DocumentTypeEditor({
               </div>
             )}
 
+            {isProtocol && (
+              <div className="space-y-1.5">
+                <label className="text-xs text-ink-muted">פרוטוקול של</label>
+                <Select value={draftScope} onValueChange={setDraftScope} dir="rtl">
+                  <SelectTrigger className="w-full">
+                    <SelectValue />
+                  </SelectTrigger>
+                  <SelectContent>
+                    {PROTOCOL_SCOPE_OPTIONS.map((opt) => (
+                      <SelectItem key={opt.value} value={opt.value}>
+                        {opt.label}
+                      </SelectItem>
+                    ))}
+                  </SelectContent>
+                </Select>
+                <p className="text-[0.65rem] text-ink-muted leading-tight">
+                  ועדת הערר → קבוצת הפרוטוקול · מקומית/מחוזית → מסמכים נלווים
+                </p>
+              </div>
+            )}
+
+            <div className="flex items-start justify-between gap-3 rounded-md border border-rule bg-parchment/40 px-3 py-2.5">
+              <div className="space-y-0.5">
+                <label
+                  htmlFor={`post-${docId}`}
+                  className="text-xs font-medium text-ink block"
+                >
+                  התקבל אחרי הדיון
+                </label>
+                <p className="text-[0.65rem] text-ink-muted leading-tight">
+                  מעביר את המסמך לקבוצת המסמכים שלאחר הדיון ומזין את בלוק-ח
+                </p>
+              </div>
+              <Switch
+                id={`post-${docId}`}
+                checked={draftPost}
+                onCheckedChange={setDraftPost}
+              />
+            </div>
+
             {patch.isError && (
               <p className="text-[0.7rem] text-danger">
                 שמירה נכשלה. נסה שוב.

web-ui/src/components/cases/documents-panel.tsx

@@ -24,6 +24,14 @@ import { apiRequest } from "@/lib/api/client";
 import { casesKeys } from "@/lib/api/cases";
 import type { CaseDetail, CaseDocument } from "@/lib/api/cases";
 import { DocumentTypeEditor } from "@/components/cases/document-type-editor";
+import { formatDateShort as formatDate } from "@/lib/format-date";
+import {
+  documentGroup,
+  DOC_GROUP_ORDER,
+  DOC_GROUP_LABELS,
+  DOC_GROUP_HINTS,
+  type DocGroup,
+} from "@/lib/doc-types";
 
 /*
  * Document list for the case detail "מסמכים" tab. Uses the real document
@@ -76,15 +84,6 @@ function StatusIcon({ status }: { status: string }) {
   }
 }
 
-function formatDate(iso: string) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL");
-  } catch {
-    return iso;
-  }
-}
-
 function filenameFromPath(path: string): string {
   const parts = path.split("/");
   return parts[parts.length - 1] || path;
@@ -232,6 +231,13 @@ function DocumentRow({
   const canPreview =
     doc.extraction_status === "completed" || doc.extraction_status === "proofread";
 
+  const meta = doc.metadata as
+    | { appraiser_side?: string; is_post_hearing?: boolean; protocol_scope?: string }
+    | undefined;
+  const isPostHearing = meta?.is_post_hearing === true;
+  const protocolScope = meta?.protocol_scope;
+  const isProtocol = doc.doc_type === "protocol";
+
   return (
     <>
       <li className="py-3 flex items-start gap-3 hover:bg-gold-wash/30 transition-colors px-2 -mx-2 rounded group">
@@ -254,15 +260,24 @@ function DocumentRow({
             {doc.created_at && <span>{formatDate(doc.created_at)}</span>}
           </div>
         </button>
+        {isProtocol && (
+          <span className="shrink-0 self-center rounded-full border border-rule bg-parchment px-2 py-0.5 text-[0.62rem] font-medium text-navy whitespace-nowrap">
+            {protocolScope === "lower" ? "ועדה מקומית" : "ועדת הערר"}
+          </span>
+        )}
+        {isPostHearing && (
+          <span className="shrink-0 self-center rounded-full border border-info/40 bg-info-bg px-2 py-0.5 text-[0.62rem] font-medium text-info whitespace-nowrap">
+            לאחר הדיון
+          </span>
+        )}
         {doc.doc_type && (
           <DocumentTypeEditor
             caseNumber={caseNumber}
             docId={doc.id}
             docType={doc.doc_type}
-            appraiserSide={
-              (doc.metadata as { appraiser_side?: string } | undefined)
-                ?.appraiser_side
-            }
+            appraiserSide={meta?.appraiser_side}
+            isPostHearing={isPostHearing}
+            protocolScope={protocolScope}
           />
         )}
         <button
@@ -314,6 +329,30 @@ function DocumentsShell({ count, children }: { count: number; children: ReactNod
   );
 }
 
+/* ── Group section header ──────────────────────────────────────────
+ * The list is split into four convenience buckets (see documentGroup).
+ * The number is the fixed 1-4 position in DOC_GROUP_ORDER so a group keeps
+ * its identity even when earlier groups are empty and hidden. */
+function GroupHeader({ group, count }: { group: DocGroup; count: number }) {
+  const index = DOC_GROUP_ORDER.indexOf(group) + 1;
+  return (
+    <div className="flex items-center gap-2.5 px-2 -mx-2 py-2 border-b border-rule-soft bg-parchment/50">
+      <span className="w-[19px] h-[19px] shrink-0 rounded-full bg-navy text-cream text-[0.66rem] font-bold flex items-center justify-center tabular-nums">
+        {index}
+      </span>
+      <span className="text-[0.8rem] font-bold text-navy shrink-0">
+        {DOC_GROUP_LABELS[group]}
+      </span>
+      <span className="shrink-0 text-[0.62rem] font-bold text-gold-deep bg-gold-wash border border-rule rounded-full px-1.5 tabular-nums">
+        {count}
+      </span>
+      <span className="text-[0.66rem] text-ink-muted font-medium truncate">
+        {DOC_GROUP_HINTS[group]}
+      </span>
+    </div>
+  );
+}
+
 export function DocumentsPanel({
   data,
 }: {
@@ -337,6 +376,16 @@ export function DocumentsPanel({
     (a, b) => statusOrder(a.extraction_status) - statusOrder(b.extraction_status),
   );
 
+  // Bucket into the four overview groups. Iterating `sorted` preserves the
+  // within-group status ordering. Grouping is display-only — see documentGroup.
+  const grouped = new Map<DocGroup, CaseDocument[]>();
+  for (const doc of sorted) {
+    const g = documentGroup(doc.doc_type, doc.metadata);
+    const bucket = grouped.get(g);
+    if (bucket) bucket.push(doc);
+    else grouped.set(g, [doc]);
+  }
+
   const done = docs.filter(
     (d) => d.extraction_status === "completed" || d.extraction_status === "proofread",
   ).length;
@@ -386,12 +435,21 @@ export function DocumentsPanel({
         </div>
       )}
 
-      <div className="max-h-[70vh] overflow-y-auto overflow-x-hidden" dir="rtl">
+      <div className="max-h-[70vh] overflow-y-auto overflow-x-hidden space-y-2" dir="rtl">
+        {DOC_GROUP_ORDER.map((group) => {
+          const groupDocs = grouped.get(group);
+          if (!groupDocs || groupDocs.length === 0) return null;
+          return (
+            <section key={group}>
+              <GroupHeader group={group} count={groupDocs.length} />
               <ul className="divide-y divide-rule" dir="rtl">
-          {sorted.map((doc) => (
+                {groupDocs.map((doc) => (
                   <DocumentRow key={doc.id} doc={doc} caseNumber={caseNumber} />
                 ))}
               </ul>
+            </section>
+          );
+        })}
       </div>
       </div>
     </DocumentsShell>

web-ui/src/components/cases/drafts-panel.tsx

@@ -25,7 +25,15 @@ import {
   useUploadFinalDecision,
   useRunFinalLearning,
   useRunFinalHalacha,
+  type ExportFile,
 } from "@/lib/api/exports";
+import {
+  usePartyClaimsSummary,
+  useGeneratePartyClaimsSummary,
+  useGenerateInterimDraft,
+  useGenerateStatus,
+} from "@/lib/api/generate";
+import { GenerateStatusChip } from "@/components/cases/generate-status-chip";
 import {
   useCaseFeedback,
   useCreateFeedback,
@@ -38,6 +46,7 @@ import {
 import { useCaseCitations } from "@/lib/api/citations";
 import { learningKeys } from "@/lib/api/learning";
 import { LearningStatusBadges } from "@/components/cases/learning-status-badges";
+import { formatDateTime, formatDateShort } from "@/lib/format-date";
 import type { CaseStatus } from "@/lib/api/cases";
 import { toast } from "sonner";
 import {
@@ -54,6 +63,10 @@ import {
   Stamp,
   Link2,
   AlertTriangle,
+  Zap,
+  ClipboardList,
+  Files,
+  ExternalLink,
 } from "lucide-react";
 
 /* Statuses at which a draft is considered ready */
@@ -72,13 +85,32 @@ function formatSize(bytes: number): string {
 }
 
 function formatDate(epoch: number): string {
-  return new Date(epoch * 1000).toLocaleDateString("he-IL", {
-    day: "numeric",
-    month: "short",
-    year: "numeric",
-    hour: "2-digit",
-    minute: "2-digit",
+  return formatDateTime(epoch * 1000);
+}
+
+/**
+ * Pick the newest export file whose name starts with `prefix`, preferring the
+ * highest v-number and falling back to the latest created_at. Used by the
+ * "הפקת מסמכים" card to surface the most recent טיוטה-טענות_הצדדים_* file (#214).
+ */
+function pickLatestVersioned(
+  files: ExportFile[] | undefined,
+  prefix: string,
+): (ExportFile & { version: number | null }) | null {
+  const matches = (files ?? []).filter((f) => f.filename.startsWith(prefix));
+  if (!matches.length) return null;
+  const withVer = matches.map((f) => {
+    // Version is the trailing integer, after either "-v" (final) or "_"
+    // (interim: טיוטה-טענות_הצדדים_{N}). Anchor to the end so the "_" inside
+    // the interim prefix isn't mistaken for the version separator.
+    const m = f.filename.match(/[_v](\d+)(?:\.\w+)?$/);
+    return { ...f, version: m ? parseInt(m[1], 10) : null };
   });
+  withVer.sort((a, b) => {
+    if (a.version != null && b.version != null) return b.version - a.version;
+    return b.created_at - a.created_at;
+  });
+  return withVer[0];
 }
 
 /* ── Main component ─────────────────────────────────── */
@@ -102,6 +134,13 @@ export function DraftsPanel({
   const uploadFinal = useUploadFinalDecision(caseNumber);
   const runLearning = useRunFinalLearning(caseNumber);
   const runHalacha = useRunFinalHalacha(caseNumber);
+  // ── Document generation (#214) — the "הפקת מסמכים" card ──
+  const { data: summary } = usePartyClaimsSummary(caseNumber);
+  const genSummary = useGeneratePartyClaimsSummary(caseNumber);
+  const genInterim = useGenerateInterimDraft(caseNumber);
+  // Server-derived background-run status (#227 ג) — survives navigation/reload.
+  const summaryStatus = useGenerateStatus(caseNumber, "party_claims_summary");
+  const interimStatus = useGenerateStatus(caseNumber, "interim_draft");
   const qc = useQueryClient();
 
   const fileRef = useRef<HTMLInputElement>(null);
@@ -134,6 +173,38 @@ export function DraftsPanel({
     return `טיוטה v${maxVer}${suffix} מוכנה לעיון`;
   })();
 
+  // ── "הפקת מסמכים" derived state (#214) ──
+  // Latest interim ("party-claims") partial draft in the exports table — the poll
+  // target for button 2. The CEO writes it as טיוטה-טענות_הצדדים_{N}.docx.
+  const latestInterim = pickLatestVersioned(exports, "טיוטה-טענות_הצדדים_");
+  const summaryReady = Boolean(summary?.markdown);
+
+  function handleGenerateSummary() {
+    genSummary.mutate(undefined, {
+      onSuccess: (d) => {
+        if (d.status === "ok") {
+          toast.success("הופעלה הפקת סיכום-המנהלים — רצה ברקע (אופוס)");
+        } else {
+          toast.warning(`לא הופעלה ההפקה: ${d.reason ?? d.error ?? d.status}`);
+        }
+      },
+      onError: () => toast.error("שגיאה בהפעלת הפקת סיכום-המנהלים"),
+    });
+  }
+
+  function handleGenerateInterim() {
+    genInterim.mutate(undefined, {
+      onSuccess: (d) => {
+        if (d.status === "ok") {
+          toast.success("הופעלה הפקת טיוטת טענות-הצדדים — רצה ברקע (אופוס)");
+        } else {
+          toast.warning(`לא הופעלה ההפקה: ${d.reason ?? d.error ?? d.status}`);
+        }
+      },
+      onError: () => toast.error("שגיאה בהפעלת הפקת הטיוטה החלקית"),
+    });
+  }
+
   function handleUpload(file: File) {
     uploadDraft.mutate(file, {
       onSuccess: (data) => {
@@ -238,34 +309,194 @@ export function DraftsPanel({
   }
 
   return (
-    <div className="space-y-6">
-      {/* ── Banner ── */}
+    // flex + per-card `order` so the visual order matches mockup 26:
+    // generate-card → banner → active-draft → drafts → final+learning → feedback.
+    <div className="flex flex-col gap-6">
+      {/* ── Card: document generation (mockup 26) — 3 deterministic "הפק" buttons ── */}
+      <section className="order-0 rounded-lg border border-rule bg-surface overflow-hidden">
+        <div className="flex items-center gap-2.5 px-4 py-3 border-b border-rule-soft bg-parchment">
+          <Zap className="w-[18px] h-[18px] text-gold-deep shrink-0" />
+          <h3 className="text-navy text-base font-semibold">הפקת מסמכים</h3>
+        </div>
+        <div className="p-4">
+          <p className="text-xs text-ink-muted leading-relaxed mb-1">
+            הפקה דטרמיניסטית בלחיצה — בלי בקשת-טקסט. כל כפתור מריץ פעולה מדויקת,
+            והקובץ מופיע בטבלת-הטיוטות / לקריאה מקומית.
+          </p>
+
+          {/* 1 — executive summary of party claims */}
+          <div className="flex items-start gap-3 py-3 border-t border-rule-soft first-of-type:border-t-0">
+            <span className="flex w-9 h-9 shrink-0 items-center justify-center rounded-lg border border-rule bg-gold-wash text-gold-deep">
+              <ClipboardList className="w-[18px] h-[18px]" />
+            </span>
+            <div className="flex-1 min-w-0">
+              <div className="text-sm font-semibold text-navy">הפק סיכום מנהלים</div>
+              <p className="text-xs text-ink-muted mt-0.5 leading-relaxed">
+                תמצית מזוקקת של טענות הצדדים לקראת הדיון — מסמך נפרד מהטיוטה.
+              </p>
+              <div className="flex flex-wrap items-center gap-2.5 mt-2">
+                <Button
+                  size="sm"
+                  onClick={handleGenerateSummary}
+                  disabled={genSummary.isPending}
+                  className="bg-navy hover:bg-navy-soft text-parchment"
+                >
+                  {genSummary.isPending ? (
+                    <>
+                      <Loader2 className="w-3.5 h-3.5 animate-spin me-1.5" />
+                      מייצר…
+                    </>
+                  ) : (
+                    "הפק"
+                  )}
+                </Button>
+                <GenerateStatusChip
+                  status={summaryStatus.data}
+                  onRetry={handleGenerateSummary}
+                  retrying={genSummary.isPending}
+                />
+                {summaryReady && (
+                  <>
+                    <Button
+                      variant="ghost"
+                      size="sm"
+                      className="h-7 px-2 text-gold-deep"
+                      onClick={() =>
+                        window.open(
+                          `/api/cases/${caseNumber}/research/party-claims-summary/download`,
+                          "_blank",
+                        )
+                      }
+                    >
+                      <ExternalLink className="w-3.5 h-3.5 me-1" />
+                      פתח
+                    </Button>
+                    <Button
+                      variant="ghost"
+                      size="sm"
+                      className="h-7 px-2 text-gold-deep"
+                      onClick={() =>
+                        window.open(
+                          `/api/cases/${caseNumber}/research/party-claims-summary/export-docx`,
+                          "_blank",
+                        )
+                      }
+                    >
+                      <Download className="w-3.5 h-3.5 me-1" />
+                      DOCX
+                    </Button>
+                  </>
+                )}
+              </div>
+            </div>
+          </div>
+
+          {/* 2 — partial draft (parties' claims, blocks ה–ט) */}
+          <div className="flex items-start gap-3 py-3 border-t border-rule-soft">
+            <span className="flex w-9 h-9 shrink-0 items-center justify-center rounded-lg border border-rule bg-gold-wash text-gold-deep">
+              <FileText className="w-[18px] h-[18px]" />
+            </span>
+            <div className="flex-1 min-w-0">
+              <div className="text-sm font-semibold text-navy">
+                הפק טיוטה של טענות הצדדים{" "}
+                <span className="font-medium text-ink-muted text-xs">
+                  (טיוטה חלקית)
+                </span>
+              </div>
+              <p className="text-xs text-ink-muted mt-0.5 leading-relaxed">
+                בלוקים ה–ט (רקע, טענות, הליכים, תכניות) — עד הדיון, בלי דיון
+                והכרעה. נשמרת בטבלת-הטיוטות.
+              </p>
+              <div className="flex flex-wrap items-center gap-2.5 mt-2">
+                <Button
+                  size="sm"
+                  onClick={handleGenerateInterim}
+                  disabled={genInterim.isPending}
+                  className="bg-navy hover:bg-navy-soft text-parchment"
+                >
+                  {genInterim.isPending ? (
+                    <>
+                      <Loader2 className="w-3.5 h-3.5 animate-spin me-1.5" />
+                      מייצר…
+                    </>
+                  ) : (
+                    "הפק"
+                  )}
+                </Button>
+                <GenerateStatusChip
+                  status={interimStatus.data}
+                  onRetry={handleGenerateInterim}
+                  retrying={genInterim.isPending}
+                />
+                {latestInterim && (
+                  <span className="text-[0.7rem] text-ink-muted">
+                    {latestInterim.version
+                      ? `v${latestInterim.version} · `
+                      : ""}
+                    {formatDate(latestInterim.created_at)}
+                  </span>
+                )}
+              </div>
+            </div>
+          </div>
+
+          {/* 3 — full decision draft (REUSES useExportDocx; moved from the banner) */}
+          <div className="flex items-start gap-3 py-3 border-t border-rule-soft">
+            <span className="flex w-9 h-9 shrink-0 items-center justify-center rounded-lg border border-rule bg-gold-wash text-gold-deep">
+              <Files className="w-[18px] h-[18px]" />
+            </span>
+            <div className="flex-1 min-w-0">
+              <div className="text-sm font-semibold text-navy">
+                הפק טיוטת החלטה מלאה
+              </div>
+              <p className="text-xs text-ink-muted mt-0.5 leading-relaxed">
+                ההחלטה המלאה — כל 12 הבלוקים, כולל דיון והכרעה. נקבעת
+                כמקור-האמת ונשמרת בטבלת-הטיוטות.
+              </p>
+              <div className="flex flex-wrap items-center gap-2.5 mt-2">
+                <Button
+                  size="sm"
+                  onClick={handleExport}
+                  disabled={exportDocx.isPending || !isDraftReady}
+                  title={!isDraftReady ? "זמין כשיש בלוקים כתובים" : undefined}
+                  className="bg-navy hover:bg-navy-soft text-parchment"
+                >
+                  {exportDocx.isPending ? (
+                    <>
+                      <Loader2 className="w-3.5 h-3.5 animate-spin me-1.5" />
+                      מייצר…
+                    </>
+                  ) : (
+                    <>
+                      <FileOutput className="w-3.5 h-3.5 me-1.5" />
+                      הפק
+                    </>
+                  )}
+                </Button>
+                {!isDraftReady && (
+                  <span className="rounded bg-warn-bg px-2 py-0.5 text-[0.65rem] text-warn">
+                    זמין כשיש בלוקים כתובים
+                  </span>
+                )}
+              </div>
+            </div>
+          </div>
+        </div>
+      </section>
+
+      {/* ── Banner — draft-ready indicator (export button moved to "הפקת מסמכים") ── */}
       {isDraftReady && (
-        <div className="flex items-center gap-3 rounded-lg border border-gold/40 bg-gold-wash px-4 py-3">
+        <div className="order-1 flex items-center gap-3 rounded-lg border border-gold/40 bg-gold-wash px-4 py-3">
           <FileText className="w-5 h-5 text-gold-deep shrink-0" />
           <span className="text-sm font-medium text-gold-deep">
             {draftLabel}
           </span>
-          <div className="me-auto" />
-          <Button
-            size="sm"
-            onClick={handleExport}
-            disabled={exportDocx.isPending}
-            className="bg-navy hover:bg-navy-soft text-parchment"
-          >
-            {exportDocx.isPending ? (
-              <Loader2 className="w-4 h-4 animate-spin me-1.5" />
-            ) : (
-              <FileOutput className="w-4 h-4 me-1.5" />
-            )}
-            הפק DOCX
-          </Button>
         </div>
       )}
 
       {/* ── Active-draft badge — the DOCX that is the current source of truth ── */}
       {activeDraft?.filename && (
-        <div className="flex items-center gap-2 text-xs text-ink-muted">
+        <div className="order-2 flex items-center gap-2 text-xs text-ink-muted">
           <span>מקור האמת:</span>
           <Badge variant="outline" className="bg-surface">
             {activeDraft.filename}
@@ -274,7 +505,7 @@ export function DraftsPanel({
       )}
 
       {/* ── Chair's signed final decision — clean upload + staged learning pipeline ── */}
-      <section className="rounded-lg border border-gold/40 bg-gold-wash/40 p-4 space-y-3">
+      <section className="order-4 rounded-lg border border-gold/40 bg-gold-wash/40 p-4 space-y-3">
         <input
           ref={finalFileRef}
           type="file"
@@ -352,13 +583,13 @@ export function DraftsPanel({
           </div>
         )}
         {hasFinal && <LearningStatusBadges caseNumber={caseNumber} />}
+        {/* precedents cited inside the signed decision — sub-section of the
+            final card (mockup 18h), collapsible inside CitationsSection itself */}
+        <CitationsSection caseNumber={caseNumber} />
       </section>
 
-      {/* ── Precedents cited inside the signed decision ── */}
-      <CitationsSection caseNumber={caseNumber} />
-
-      {/* ── Exports list ── */}
-      <section>
+      {/* ── Card: draft files (mockup 18h) ── */}
+      <section className="order-3 rounded-lg border border-rule bg-surface p-4">
         <div className="flex items-center justify-between mb-3">
           <h3 className="text-navy text-base">קבצי טיוטה</h3>
           <div className="flex items-center gap-2">
@@ -401,9 +632,9 @@ export function DraftsPanel({
             <table className="w-full text-sm">
               <thead>
                 <tr className="bg-rule-soft/40 text-ink-muted text-[0.75rem]">
-                  <th className="text-start px-4 py-2 font-medium">File</th>
-                  <th className="text-start px-4 py-2 font-medium">Size</th>
-                  <th className="text-start px-4 py-2 font-medium">Date</th>
+                  <th className="text-start px-4 py-2 font-medium">קובץ</th>
+                  <th className="text-start px-4 py-2 font-medium">גודל</th>
+                  <th className="text-start px-4 py-2 font-medium">תאריך</th>
                   <th className="px-4 py-2" />
                 </tr>
               </thead>
@@ -526,8 +757,8 @@ export function DraftsPanel({
         </Dialog>
       </section>
 
-      {/* ── Chair feedback ── */}
-      <section>
+      {/* ── Card: chair feedback (mockup 18h) ── */}
+      <section className="order-5 rounded-lg border border-rule bg-surface p-4">
         <div className="flex items-center justify-between mb-3">
           <h3 className="text-navy text-base">
             הערות יו״ר
@@ -568,9 +799,7 @@ export function DraftsPanel({
                     </Badge>
                   )}
                   <span className="text-[0.65rem] text-ink-muted me-auto">
-                    {fb.created_at
-                      ? new Date(fb.created_at).toLocaleDateString("he-IL")
-                      : ""}
+                    {fb.created_at ? formatDateShort(fb.created_at) : ""}
                   </span>
                 </div>
                 <p className="text-sm leading-relaxed">{fb.feedback_text}</p>

web-ui/src/components/cases/generate-status-chip.tsx

@@ -0,0 +1,127 @@
+"use client";
+
+import { useEffect, useState } from "react";
+import { Button } from "@/components/ui/button";
+import { CheckCircle2, Clock, Loader2, RotateCw, XCircle } from "lucide-react";
+import type { GenerateStatus } from "@/lib/api/generate";
+
+/**
+ * Background-run status chip for the "הפקת מסמכים" generation actions (#227 ג).
+ *
+ * The chip is driven entirely by the server-derived {@link GenerateStatus} — it
+ * holds no run state of its own, so it renders correctly after the chair
+ * navigates away and back. The running timer is anchored to the server
+ * `started_at`, not a click-time counter, so the elapsed value is right on
+ * return too.
+ */
+
+function useElapsedSeconds(startedAt: string | null, active: boolean): number | null {
+  const [now, setNow] = useState(() => Date.now());
+  useEffect(() => {
+    if (!active) return;
+    const id = setInterval(() => setNow(Date.now()), 1_000);
+    return () => clearInterval(id);
+  }, [active]);
+  if (!startedAt) return null;
+  const start = Date.parse(startedAt);
+  if (Number.isNaN(start)) return null;
+  return Math.max(0, Math.floor((now - start) / 1000));
+}
+
+function fmtDuration(sec: number): string {
+  const m = Math.floor(sec / 60);
+  const s = sec % 60;
+  if (m >= 60) {
+    const h = Math.floor(m / 60);
+    return `${h}:${String(m % 60).padStart(2, "0")}:${String(s).padStart(2, "0")}`;
+  }
+  return `${m}:${String(s).padStart(2, "0")}`;
+}
+
+function relativeTime(iso: string | null): string {
+  if (!iso) return "";
+  const t = Date.parse(iso);
+  if (Number.isNaN(t)) return "";
+  const sec = Math.max(0, Math.floor((Date.now() - t) / 1000));
+  if (sec < 60) return "הרגע";
+  const min = Math.floor(sec / 60);
+  if (min < 60) return `לפני ${min} דק׳`;
+  const hr = Math.floor(min / 60);
+  if (hr < 24) return `לפני ${hr} שע׳`;
+  return `לפני ${Math.floor(hr / 24)} ימים`;
+}
+
+const CHIP_BASE =
+  "inline-flex items-center gap-1.5 rounded-full border px-2.5 py-1 text-xs font-semibold whitespace-nowrap";
+
+export function GenerateStatusChip({
+  status,
+  onRetry,
+  retrying,
+}: {
+  status: GenerateStatus | undefined;
+  onRetry?: () => void;
+  retrying?: boolean;
+}) {
+  const state = status?.state ?? "idle";
+  const elapsed = useElapsedSeconds(status?.started_at ?? null, state === "running");
+
+  if (!status || state === "idle") return null;
+
+  if (state === "queued") {
+    return (
+      <span className={`${CHIP_BASE} bg-rule-soft text-ink-soft border-rule`}>
+        <Clock className="size-3.5" aria-hidden />
+        בתור…
+      </span>
+    );
+  }
+
+  if (state === "running") {
+    return (
+      <span className={`${CHIP_BASE} bg-info-bg text-info border-info/35`}>
+        <Loader2 className="size-3.5 animate-spin" aria-hidden />
+        רץ ברקע
+        {elapsed !== null && (
+          <span className="tabular-nums">· {fmtDuration(elapsed)}</span>
+        )}
+      </span>
+    );
+  }
+
+  if (state === "done") {
+    const rel = relativeTime(status.finished_at);
+    return (
+      <span className={`${CHIP_BASE} bg-success-bg text-success border-success/35`}>
+        <CheckCircle2 className="size-3.5" aria-hidden />
+        הושלם{rel && ` · ${rel}`}
+      </span>
+    );
+  }
+
+  // failed
+  return (
+    <span className="inline-flex items-center gap-2">
+      <span className={`${CHIP_BASE} bg-danger-bg text-danger border-danger/35`}>
+        <XCircle className="size-3.5" aria-hidden />
+        {status.detail || "נכשל"}
+      </span>
+      {onRetry && (
+        <Button
+          variant="ghost"
+          size="sm"
+          className="h-7 px-2 text-gold-deep"
+          disabled={retrying}
+          onClick={onRetry}
+        >
+          {retrying ? (
+            <Loader2 className="size-3.5 animate-spin me-1" />
+          ) : (
+            <RotateCw className="size-3.5 me-1" />
+          )}
+          נסה שוב
+        </Button>
+      )}
+    </span>
+  );
+}

web-ui/src/components/cases/hearing-changes-panel.tsx

@@ -0,0 +1,360 @@
+"use client";
+
+import { useMemo, useState } from "react";
+import { Badge } from "@/components/ui/badge";
+import { Button } from "@/components/ui/button";
+import { Skeleton } from "@/components/ui/skeleton";
+import {
+  ArrowDown,
+  ArrowUp,
+  CalendarDays,
+  HelpCircle,
+  Link2,
+  Loader2,
+  Plus,
+  RotateCw,
+  Sparkles,
+} from "lucide-react";
+import { toast } from "sonner";
+import {
+  CHANGE_LABELS_HE,
+  PROTOCOL_PARTY_LABELS_HE,
+  PROTOCOL_PARTY_ORDER,
+  useAnalyzeProtocol,
+  useProtocolAnalysis,
+  type ProtocolAnalysisRow,
+  type ProtocolChangeType,
+  type ProtocolPartyRole,
+} from "@/lib/api/protocol-analysis";
+
+const PARTY_MARK_TONE: Record<ProtocolPartyRole, string> = {
+  appellant: "bg-info",
+  respondent: "bg-gold-deep",
+  committee: "bg-success",
+  permit_applicant: "bg-warn",
+  unknown: "bg-ink-muted",
+  "": "bg-ink-muted",
+};
+
+const CHANGE_BADGE_TONE: Record<ProtocolChangeType, string> = {
+  strengthened: "bg-gold-wash text-gold-deep border-gold/40",
+  newly_raised: "bg-emerald-50 text-emerald-900 border-emerald-200",
+  dropped: "bg-rule-soft text-ink-soft border-rule",
+};
+
+const CHANGE_ACCENT: Record<ProtocolChangeType, string> = {
+  strengthened: "border-s-gold",
+  newly_raised: "border-s-emerald-500",
+  dropped: "border-s-rule",
+};
+
+function ChangeIcon({ type }: { type: ProtocolChangeType }) {
+  if (type === "strengthened") return <ArrowUp className="size-3.5" aria-hidden />;
+  if (type === "newly_raised") return <Plus className="size-3.5" aria-hidden />;
+  return <ArrowDown className="size-3.5" aria-hidden />;
+}
+
+type FilterKey = "all" | ProtocolChangeType;
+
+function AttendeeColumn({
+  label,
+  dotTone,
+  names,
+}: {
+  label: string;
+  dotTone: string;
+  names: string[];
+}) {
+  if (!names.length) return null;
+  return (
+    <div className="min-w-[210px] flex-1">
+      <div className="text-gold-deep mb-1.5 flex items-center gap-1.5 text-[0.7rem] font-bold">
+        <span className={`size-1.5 flex-none rounded-full ${dotTone}`} aria-hidden />
+        {label}
+      </div>
+      <div className="text-ink-soft text-xs leading-relaxed">
+        {names.join(" · ")}
+      </div>
+    </div>
+  );
+}
+
+function DevelopmentCard({ row }: { row: ProtocolAnalysisRow }) {
+  return (
+    <div
+      className={`bg-surface border-rule overflow-hidden rounded-lg border border-s-4 shadow-sm ${CHANGE_ACCENT[row.change_type]}`}
+    >
+      <div className="space-y-2.5 px-4 py-3.5">
+        <div className="flex flex-wrap items-start gap-2.5">
+          <Badge
+            variant="outline"
+            className={`${CHANGE_BADGE_TONE[row.change_type]} flex-none gap-1 text-[0.72rem] font-bold`}
+          >
+            <ChangeIcon type={row.change_type} />
+            {CHANGE_LABELS_HE[row.change_type]}
+          </Badge>
+          <span className="text-navy flex-1 text-sm font-bold leading-snug">
+            {row.argument_title}
+          </span>
+        </div>
+
+        {row.summary && (
+          <p className="text-ink-soft text-[0.82rem] leading-relaxed">
+            {row.summary}
+          </p>
+        )}
+
+        {row.sharpened_question && (
+          <div className="bg-parchment border-rule-soft rounded-md border px-3 py-2">
+            <div className="text-gold-deep mb-1 flex items-center gap-1.5 text-[0.66rem] font-bold">
+              <HelpCircle className="size-3" aria-hidden />
+              השאלה שהתחדדה
+            </div>
+            <div className="text-ink text-xs leading-relaxed">
+              {row.sharpened_question}
+            </div>
+          </div>
+        )}
+
+        {row.evidence_quote && (
+          <blockquote className="border-rule text-ink-muted border-s-[3px] px-3 text-xs italic leading-relaxed">
+            &ldquo;{row.evidence_quote}&rdquo;
+            <span className="text-ink-muted mt-1 block text-[0.68rem] not-italic font-semibold">
+              — מתוך פרוטוקול הדיון
+            </span>
+          </blockquote>
+        )}
+
+        <div className="flex flex-wrap items-center gap-2 pt-0.5">
+          {row.change_type === "newly_raised" ? (
+            <span className="text-emerald-800 inline-flex items-center gap-1.5 text-[0.72rem] font-semibold">
+              <Sparkles className="size-3.5" aria-hidden />
+              אין מקבילה בכתב — עלה לראשונה בדיון
+            </span>
+          ) : row.argument_id ? (
+            <span className="text-ink-muted inline-flex items-center gap-1.5 text-[0.72rem] font-semibold">
+              <Link2 className="size-3.5" aria-hidden />
+              מבוסס על טיעון כתוב קיים
+            </span>
+          ) : null}
+        </div>
+      </div>
+    </div>
+  );
+}
+
+type HearingChangesPanelProps = {
+  caseNumber: string;
+};
+
+export function HearingChangesPanel({ caseNumber }: HearingChangesPanelProps) {
+  const { data, isPending, isError, error } = useProtocolAnalysis(caseNumber);
+  const analyze = useAnalyzeProtocol(caseNumber);
+  const [filter, setFilter] = useState<FilterKey>("all");
+
+  const handleAnalyze = () => {
+    analyze.mutate(undefined, {
+      onSuccess: (res) => {
+        if (res.status === "queued") {
+          toast.success("נשלח למנתח המשפטי — הניתוח ירוץ ברקע; רענן בעוד כמה דקות.");
+        } else {
+          toast.warning(
+            `לא ניתן להריץ אוטומטית (${res.reason}). ניתן להריץ ידנית מ-Claude Code.`,
+          );
+        }
+      },
+      onError: (e) => toast.error(`שגיאה: ${(e as Error).message}`),
+    });
+  };
+
+  const header = data?.header;
+  const counts = data?.by_change ?? {};
+  const strengthened = counts.strengthened ?? 0;
+  const newlyRaised = counts.newly_raised ?? 0;
+  const dropped = counts.dropped ?? 0;
+
+  const sections = useMemo(() => {
+    if (!data) return [];
+    return PROTOCOL_PARTY_ORDER.map((role) => {
+      const rows = (data.by_party[role] ?? []).filter(
+        (r) => filter === "all" || r.change_type === filter,
+      );
+      return { role, rows };
+    }).filter((s) => s.rows.length > 0);
+  }, [data, filter]);
+
+  const runButton = (
+    <Button
+      variant="outline"
+      size="sm"
+      disabled={analyze.isPending}
+      onClick={handleAnalyze}
+    >
+      {analyze.isPending ? (
+        <Loader2 className="me-1.5 size-3.5 animate-spin" />
+      ) : (
+        <RotateCw className="me-1.5 size-3.5" />
+      )}
+      {data?.total ? "נתח מחדש" : "נתח את פרוטוקול הדיון"}
+    </Button>
+  );
+
+  return (
+    <div className="space-y-4">
+      {isPending ? (
+        <div className="space-y-2">
+          <Skeleton className="h-20 w-full" />
+          <Skeleton className="h-24 w-full" />
+        </div>
+      ) : isError ? (
+        <p className="text-danger text-sm">
+          שגיאה בטעינת ניתוח-הפרוטוקול: {(error as Error).message}
+        </p>
+      ) : !data?.total ? (
+        <div className="bg-surface border-rule rounded-lg border border-dashed px-6 py-8 text-center">
+          <h3 className="text-navy text-sm font-bold">טרם נותח פרוטוקול</h3>
+          <p className="text-ink-muted mx-auto mt-1.5 max-w-md text-xs leading-relaxed">
+            כשיש בתיק פרוטוקול של ועדת-הערר וטיעונים מאוגדים, הניתוח מזהה מה
+            התחזק ומה נטען לראשונה בדיון. ההרצה נשלחת למנתח המשפטי ורצה ברקע.
+          </p>
+          <div className="mt-4 flex justify-center">{runButton}</div>
+        </div>
+      ) : (
+        <>
+          <div className="flex justify-end">{runButton}</div>
+          {/* hearing header strip */}
+          {header && (
+            <div className="bg-parchment border-rule rounded-lg border px-5 py-4 shadow-sm">
+              <div className="flex flex-wrap items-center gap-3">
+                <span className="bg-gold-wash border-rule flex size-9 flex-none items-center justify-center rounded-lg border">
+                  <CalendarDays className="text-gold-deep size-5" aria-hidden />
+                </span>
+                <div className="flex-1">
+                  <div className="text-navy text-[0.95rem] font-bold">
+                    {header.protocol_title || "פרוטוקול דיון"}
+                  </div>
+                  <div className="text-ink-muted mt-0.5 text-xs">
+                    {header.hearing_date
+                      ? `תאריך דיון: ${header.hearing_date}`
+                      : "תאריך דיון לא זוהה"}
+                    {header.panel_members.length > 0 &&
+                      ` · מותב: ${header.panel_members.join(", ")}`}
+                  </div>
+                </div>
+                <Badge
+                  variant="outline"
+                  className="bg-info-bg text-info border-info/30 flex-none rounded-full text-[0.72rem] font-semibold"
+                >
+                  פרוטוקול ועדת-הערר
+                </Badge>
+              </div>
+              {(header.appellants_present.length > 0 ||
+                header.respondents_present.length > 0) && (
+                <div className="border-rule-soft mt-3 flex flex-wrap gap-6 border-t pt-3">
+                  <AttendeeColumn
+                    label="נכחו מטעם העוררים"
+                    dotTone="bg-info"
+                    names={header.appellants_present}
+                  />
+                  <AttendeeColumn
+                    label="נכחו מטעם המשיבים"
+                    dotTone="bg-gold-deep"
+                    names={header.respondents_present}
+                  />
+                </div>
+              )}
+            </div>
+          )}
+
+          {/* filter pills */}
+          <div className="flex flex-wrap items-center gap-2">
+            <FilterPill
+              active={filter === "all"}
+              onClick={() => setFilter("all")}
+              label="הכל"
+              count={data.total}
+            />
+            {strengthened > 0 && (
+              <FilterPill
+                active={filter === "strengthened"}
+                onClick={() => setFilter("strengthened")}
+                label="התחזקו"
+                count={strengthened}
+                icon={<ArrowUp className="size-3.5" aria-hidden />}
+              />
+            )}
+            {newlyRaised > 0 && (
+              <FilterPill
+                active={filter === "newly_raised"}
+                onClick={() => setFilter("newly_raised")}
+                label="נטענו לראשונה"
+                count={newlyRaised}
+                icon={<Plus className="size-3.5" aria-hidden />}
+              />
+            )}
+            {dropped > 0 && (
+              <FilterPill
+                active={filter === "dropped"}
+                onClick={() => setFilter("dropped")}
+                label="נזנחו"
+                count={dropped}
+                icon={<ArrowDown className="size-3.5" aria-hidden />}
+              />
+            )}
+          </div>
+
+          {/* per-party sections */}
+          {sections.map(({ role, rows }) => (
+            <div key={role || "unassigned"} className="space-y-2.5">
+              <div className="flex items-center gap-2.5">
+                <span
+                  className={`h-5 w-1.5 flex-none rounded ${PARTY_MARK_TONE[role]}`}
+                  aria-hidden
+                />
+                <span className="text-navy text-[0.95rem] font-bold">
+                  {PROTOCOL_PARTY_LABELS_HE[role]}
+                </span>
+                <span className="text-ink-muted text-xs">{rows.length}</span>
+              </div>
+              <div className="space-y-2.5">
+                {rows.map((r) => (
+                  <DevelopmentCard key={r.id} row={r} />
+                ))}
+              </div>
+            </div>
+          ))}
+        </>
+      )}
+    </div>
+  );
+}
+
+function FilterPill({
+  active,
+  onClick,
+  label,
+  count,
+  icon,
+}: {
+  active: boolean;
+  onClick: () => void;
+  label: string;
+  count: number;
+  icon?: React.ReactNode;
+}) {
+  return (
+    <button
+      type="button"
+      onClick={onClick}
+      className={`inline-flex items-center gap-1.5 rounded-full border px-3.5 py-1.5 text-xs font-semibold transition-colors ${
+        active
+          ? "bg-navy border-navy text-white"
+          : "bg-surface border-rule text-ink-soft hover:bg-parchment"
+      }`}
+    >
+      {icon}
+      {label}
+      <span className="font-bold">{count}</span>
+    </button>
+  );
+}

web-ui/src/components/cases/legal-arguments-panel.tsx

@@ -36,6 +36,23 @@ const PRIORITY_BADGE_TONE: Record<LegalArgumentPriority, string> = {
   relief: "bg-emerald-50 text-emerald-900 border-emerald-200",
 };
 
+// Per-party colour mark + one-line role description (mockup 25b party trigger).
+const PARTY_MARK_TONE: Record<LegalArgumentParty, string> = {
+  appellant: "bg-info",
+  respondent: "bg-gold-deep",
+  committee: "bg-success",
+  permit_applicant: "bg-warn",
+  unknown: "bg-ink-muted",
+};
+
+const PARTY_SUB_HE: Record<LegalArgumentParty, string> = {
+  appellant: "הצד היוזם של הערר",
+  respondent: "הצד המשיב לערר",
+  committee: "עמדת הגורם המאשר",
+  permit_applicant: "צד שלישי שזכותו עשויה להיפגע",
+  unknown: "צד שלא זוהה אוטומטית",
+};
+
 function groupByPriority(
   args: LegalArgument[],
 ): Record<LegalArgumentPriority, LegalArgument[]> {
@@ -55,22 +72,57 @@ function groupByPriority(
 }
 
 type PartySectionProps = {
-  party: LegalArgumentParty;
   args: LegalArgument[];
 };
 
-function PartySection({ party, args }: PartySectionProps) {
-  const grouped = useMemo(() => groupByPriority(args), [args]);
+/**
+ * Inner body of a single party. For a multi-litigant side (respondent /
+ * permit_applicant) whose arguments carry distinct ``party_name`` brief labels,
+ * we sub-group by brief first — so opposing joint briefs (e.g. "משיבות 2-3" vs
+ * "משיבים 4-6") read as separate positions rather than one merged list (#224).
+ * Single-voice sides (one empty brief) render the priority groups directly.
+ */
+function PartySection({ args }: PartySectionProps) {
+  const briefs = useMemo(() => {
+    const byBrief = new Map<string, LegalArgument[]>();
+    for (const a of args) {
+      const key = a.party_name?.trim() ?? "";
+      const list = byBrief.get(key);
+      if (list) list.push(a);
+      else byBrief.set(key, [a]);
+    }
+    return [...byBrief.entries()];
+  }, [args]);
+
+  // No real per-brief split (single, unnamed group) → render groups directly.
+  if (briefs.length <= 1) {
+    return <PriorityGroups args={briefs[0]?.[1] ?? args} />;
+  }
+
   return (
-    <div className="space-y-3">
-      <div className="flex items-baseline justify-between border-b border-rule pb-2">
-        <h3 className="text-navy text-base font-semibold">
-          {PARTY_LABELS_HE[party] ?? party}
-        </h3>
+    <div className="space-y-5">
+      {briefs.map(([brief, briefArgs]) => (
+        <div key={brief || "—"} className="space-y-2">
+          <div className="flex items-center gap-2 border-s-2 border-gold-deep ps-2">
+            <span className="text-navy text-sm font-semibold">
+              {brief || "ללא שיוך לכתב"}
+            </span>
             <span className="text-ink-muted text-xs">
-          {args.length} טיעונים
+              {briefArgs.length} טיעונים
             </span>
           </div>
+          <PriorityGroups args={briefArgs} />
+        </div>
+      ))}
+    </div>
+  );
+}
+
+/** Priority-grouped argument accordion for one coherent set of arguments. */
+function PriorityGroups({ args }: PartySectionProps) {
+  const grouped = useMemo(() => groupByPriority(args), [args]);
+  return (
+    <div className="space-y-4">
       {PRIORITY_ORDER.map((priority) => {
         const list = grouped[priority];
         if (!list?.length) return null;
@@ -253,15 +305,47 @@ export function LegalArgumentsPanel({ caseNumber }: LegalArgumentsPanelProps) {
           החישוב למנתח המשפטי.
         </p>
       ) : (
-        <div className="space-y-6">
-          {parties.map((party) => (
-            <PartySection
+        // Top-level party accordion — each party is its own collapsible section,
+        // closed by default (empty defaultValue). The count badge on every
+        // trigger shows the scope without expanding (mockup 25b).
+        <Accordion type="multiple" className="space-y-3.5">
+          {parties.map((party) => {
+            const args = data.by_party[party] ?? [];
+            return (
+              <AccordionItem
                 key={party}
-              party={party}
-              args={data.by_party[party] ?? []}
+                value={party}
+                className="overflow-hidden rounded-lg border border-rule bg-surface shadow-sm"
+              >
+                <AccordionTrigger className="px-5 py-4 hover:no-underline data-[state=open]:border-b data-[state=open]:border-rule-soft data-[state=open]:bg-parchment">
+                  <span className="flex flex-1 items-center gap-3.5">
+                    <span
+                      className={`h-8 w-1.5 flex-none rounded ${PARTY_MARK_TONE[party]}`}
+                      aria-hidden
                     />
-          ))}
-        </div>
+                    <span className="flex flex-1 flex-col items-start gap-0.5">
+                      <span className="text-navy text-base font-bold leading-tight">
+                        {PARTY_LABELS_HE[party] ?? party}
+                      </span>
+                      <span className="text-ink-muted text-xs">
+                        {PARTY_SUB_HE[party]}
+                      </span>
+                    </span>
+                    <Badge
+                      variant="outline"
+                      className="bg-gold-wash text-gold-deep border-rule rounded-full px-3 py-0.5 text-xs font-bold whitespace-nowrap"
+                    >
+                      {args.length} טיעונים
+                    </Badge>
+                  </span>
+                </AccordionTrigger>
+                <AccordionContent className="px-5 pb-4 pt-4">
+                  <PartySection args={args} />
+                </AccordionContent>
+              </AccordionItem>
+            );
+          })}
+        </Accordion>
       )}
     </div>
   );

web-ui/src/components/cases/positions-panel.tsx

@@ -0,0 +1,299 @@
+"use client";
+
+import { useRef, useState } from "react";
+import { Card, CardContent } from "@/components/ui/card";
+import { Button } from "@/components/ui/button";
+import { Skeleton } from "@/components/ui/skeleton";
+import { SubsectionCard } from "@/components/compose/subsection-card";
+import { PrecedentsSection } from "@/components/compose/precedents-section";
+import { Markdown } from "@/components/ui/markdown";
+import { useCase } from "@/lib/api/cases";
+import { useResearchAnalysis } from "@/lib/api/research";
+import { useCasePrecedents } from "@/lib/api/precedents";
+
+/*
+ * PositionsPanel — the chair's editing surface for the analyst's threshold
+ * claims + issues (chair positions feed block י). Extracted verbatim from the
+ * former /compose "עמדות וטענות" tab when the decision editor was merged into
+ * the main case page (X17 #3). Lives inside the "טיעונים ועמדות" tab above the
+ * collapsible by-party aggregated arguments. The 12-block editor + citation
+ * verification moved to their own top-level tabs; /compose was deleted.
+ */
+
+function ProseSection({ title, content }: { title: string; content?: string }) {
+  if (!content?.trim()) return null;
+  return (
+    <section className="space-y-2">
+      <h3 className="text-[0.78rem] uppercase tracking-[0.08em] text-gold-deep font-semibold">
+        {title}
+      </h3>
+      <Markdown content={content.trim()} />
+    </section>
+  );
+}
+
+// ── "ייצוא ועדכון הניתוח" rail card — round-trips the analysis-and-research.md
+// research analysis: export DOCX, upload an updated version, download the raw MD.
+// (Post-final learning status lives on the drafts tab, not here — #226.) ──────
+function FinishRail({
+  caseNumber,
+  hasAnalysis,
+  onUploaded,
+}: {
+  caseNumber: string;
+  hasAnalysis: boolean;
+  onUploaded: () => void;
+}) {
+  const fileRef = useRef<HTMLInputElement>(null);
+  const [uploading, setUploading] = useState(false);
+  const [uploadMsg, setUploadMsg] = useState<{ ok: boolean; text: string } | null>(null);
+
+  async function handleUpload(file: File) {
+    setUploading(true);
+    setUploadMsg(null);
+    try {
+      const form = new FormData();
+      form.append("file", file);
+      const res = await fetch(`/api/cases/${caseNumber}/research/analysis/upload`, {
+        method: "PUT",
+        body: form,
+      });
+      const data = await res.json();
+      if (!res.ok) {
+        setUploadMsg({ ok: false, text: data.detail || "שגיאה בהעלאה" });
+        return;
+      }
+      setUploadMsg({
+        ok: true,
+        text: `הקובץ הועלה — ${data.sections.threshold_claims} טענות סף, ${data.sections.issues} סוגיות`,
+      });
+      onUploaded();
+    } catch {
+      setUploadMsg({ ok: false, text: "שגיאת רשת" });
+    } finally {
+      setUploading(false);
+      if (fileRef.current) fileRef.current.value = "";
+    }
+  }
+
+  return (
+    <Card className="bg-surface border-rule shadow-sm h-full">
+      <CardContent className="px-4 py-4">
+        <h3 className="text-navy text-[0.9rem] font-semibold mb-1">ייצוא ועדכון הניתוח</h3>
+        <p className="text-[0.72rem] text-ink-muted mb-3">
+          פעולות על קובץ ניתוח-המחקר (analysis-and-research.md).
+        </p>
+
+        <input
+          ref={fileRef}
+          type="file"
+          accept=".md"
+          className="hidden"
+          onChange={(e) => {
+            const f = e.target.files?.[0];
+            if (f) handleUpload(f);
+          }}
+        />
+
+        <div className="space-y-2">
+          {hasAnalysis && (
+            <Button
+              variant="outline"
+              className="w-full justify-center"
+              onClick={() => {
+                const a = document.createElement("a");
+                a.href = `/api/cases/${caseNumber}/research/analysis/export-docx`;
+                a.click();
+              }}
+            >
+              ייצוא DOCX
+            </Button>
+          )}
+          <Button
+            className="w-full justify-center bg-gold text-white hover:bg-gold-deep"
+            disabled={uploading}
+            onClick={() => fileRef.current?.click()}
+          >
+            {uploading ? "מעלה…" : "העלאת ניתוח מעודכן"}
+          </Button>
+          {hasAnalysis && (
+            <Button
+              variant="outline"
+              className="w-full justify-center"
+              onClick={() => {
+                const a = document.createElement("a");
+                a.href = `/api/cases/${caseNumber}/research/analysis/download`;
+                a.download = `analysis-${caseNumber}.md`;
+                a.click();
+              }}
+            >
+              הורד ניתוח (MD)
+            </Button>
+          )}
+        </div>
+
+        {uploadMsg && (
+          <p className={`text-xs mt-2 ${uploadMsg.ok ? "text-success" : "text-danger"}`}>
+            {uploadMsg.text}
+          </p>
+        )}
+      </CardContent>
+    </Card>
+  );
+}
+
+export function PositionsPanel({ caseNumber }: { caseNumber: string }) {
+  const caseQuery = useCase(caseNumber);
+  const analysis = useResearchAnalysis(caseNumber);
+  const precedentsQuery = useCasePrecedents(caseNumber);
+
+  const allPrecedents = precedentsQuery.data ?? [];
+  const caseLevelPrecedents = allPrecedents.filter((p) => p.section_id === null);
+  const precedentsBySection = new Map<string, typeof allPrecedents>();
+  for (const p of allPrecedents) {
+    if (p.section_id) {
+      const existing = precedentsBySection.get(p.section_id) ?? [];
+      existing.push(p);
+      precedentsBySection.set(p.section_id, existing);
+    }
+  }
+  const practiceArea = caseQuery.data?.practice_area ?? null;
+
+  const isNotFound =
+    analysis.error instanceof Error &&
+    /404|לא נמצא|טרם בוצע/.test(analysis.error.message);
+
+  return (
+    <div className="space-y-4">
+      {analysis.isPending ? (
+        <Card className="bg-surface border-rule shadow-sm">
+          <CardContent className="px-6 py-5 space-y-3">
+            <Skeleton className="h-6 w-48" />
+            <Skeleton className="h-4 w-96" />
+            <Skeleton className="h-32 w-full" />
+          </CardContent>
+        </Card>
+      ) : isNotFound ? (
+        <Card className="bg-surface border-rule shadow-sm">
+          <CardContent className="px-6 py-12 text-center space-y-3">
+            <div className="text-gold text-3xl" aria-hidden>❦</div>
+            <h2 className="text-navy text-lg mb-0">
+              טרם בוצע ניתוח משפטי לתיק זה
+            </h2>
+            <p className="text-ink-muted text-sm max-w-md mx-auto">
+              לאחר שקובץ <code>analysis-and-research.md</code> ייווצר, תוכלי
+              לערוך כאן את עמדת הוועדה לכל טענת סף וסוגיה.
+            </p>
+          </CardContent>
+        </Card>
+      ) : analysis.error ? (
+        <Card className="bg-danger-bg border-danger/40">
+          <CardContent className="px-6 py-5 text-center">
+            <p className="text-danger">{analysis.error.message}</p>
+          </CardContent>
+        </Card>
+      ) : analysis.data ? (
+        <div className="space-y-6">
+          {analysis.data.threshold_claims &&
+            analysis.data.threshold_claims.length > 0 && (
+              <div className="space-y-3">
+                <div className="flex items-center gap-2">
+                  <h2 className="text-navy text-lg font-semibold mb-0">טענות סף</h2>
+                  <span className="text-[0.72rem] rounded-full bg-gold-wash text-gold-deep px-2 py-0.5 border border-gold/40 tabular-nums">
+                    {analysis.data.threshold_claims.length}
+                  </span>
+                </div>
+                <div className="grid gap-3 lg:grid-cols-2 items-start">
+                  {analysis.data.threshold_claims.map((tc) => (
+                    <SubsectionCard
+                      key={tc.id}
+                      caseNumber={caseNumber}
+                      item={tc}
+                      precedents={precedentsBySection.get(tc.id) ?? []}
+                      practiceArea={practiceArea}
+                    />
+                  ))}
+                </div>
+              </div>
+            )}
+
+          {analysis.data.issues && analysis.data.issues.length > 0 && (
+            <div className="space-y-3">
+              <div className="flex items-center gap-2">
+                <h2 className="text-navy text-lg font-semibold mb-0">סוגיות להכרעה</h2>
+                <span className="text-[0.72rem] rounded-full bg-gold-wash text-gold-deep px-2 py-0.5 border border-gold/40 tabular-nums">
+                  {analysis.data.issues.length}
+                </span>
+              </div>
+              <div className="grid gap-3 lg:grid-cols-2 items-start">
+                {analysis.data.issues.map((iss) => (
+                  <SubsectionCard
+                    key={iss.id}
+                    caseNumber={caseNumber}
+                    item={iss}
+                    precedents={precedentsBySection.get(iss.id) ?? []}
+                    practiceArea={practiceArea}
+                  />
+                ))}
+              </div>
+            </div>
+          )}
+
+          {!analysis.data.threshold_claims?.length &&
+            !analysis.data.issues?.length && (
+              <Card className="bg-surface border-rule">
+                <CardContent className="px-6 py-10 text-center text-ink-muted">
+                  לא נמצאו טענות סף או סוגיות בניתוח זה.
+                </CardContent>
+              </Card>
+            )}
+
+          <Card className="bg-surface border-rule shadow-sm">
+            <CardContent className="px-6 py-5 space-y-5">
+              <h2 className="text-navy text-lg font-semibold mb-0">רקע לניתוח</h2>
+              <ProseSection title="צד מיוצג" content={analysis.data.represented_party} />
+              <ProseSection title="רקע דיוני" content={analysis.data.procedural_background} />
+              <ProseSection title="עובדות מוסכמות" content={analysis.data.agreed_facts} />
+              <ProseSection title="עובדות במחלוקת" content={analysis.data.disputed_facts} />
+            </CardContent>
+          </Card>
+
+          {analysis.data.conclusions?.trim() && (
+            <Card className="bg-gold-wash border-gold/40 shadow-sm">
+              <CardContent className="px-6 py-5 space-y-3">
+                <h2 className="text-gold-deep text-lg font-semibold mb-0">מסקנות</h2>
+                <Markdown content={analysis.data.conclusions.trim()} />
+              </CardContent>
+            </Card>
+          )}
+        </div>
+      ) : null}
+
+      {/* case-level פסיקה מצורפת + analysis export/update (INV-IA3: not removed).
+          items-stretch keeps both cards the same height (#226). */}
+      <div className="grid gap-4 lg:grid-cols-2 items-stretch pt-2">
+        <Card className="bg-surface border-rule shadow-sm h-full">
+          <CardContent className="px-5 py-4">
+            <h3 className="text-navy text-[0.9rem] font-semibold mb-1">פסיקה מצורפת</h3>
+            <p className="text-[0.72rem] text-ink-muted mb-3">
+              ציטוטים התומכים בעמדה באופן רוחבי — ישולבו בפתיחת בלוק י (דיון).
+            </p>
+            <PrecedentsSection
+              caseNumber={caseNumber}
+              sectionId={null}
+              precedents={caseLevelPrecedents}
+              practiceArea={practiceArea}
+              emptyHelperText="עדיין לא צורפה פסיקה כללית לתיק"
+            />
+          </CardContent>
+        </Card>
+
+        <FinishRail
+          caseNumber={caseNumber}
+          hasAnalysis={!!analysis.data}
+          onUploaded={() => analysis.refetch()}
+        />
+      </div>
+    </div>
+  );
+}

web-ui/src/components/cases/status-badge.tsx

@@ -17,6 +17,8 @@ const STATUS_ICONS: Record<CaseStatus, LucideIcon> = {
   new:                FilePlus2,
   processing:         Loader2,
   documents_ready:    FileCheck,
+  analyst_verified:   SearchCheck,
+  research_complete:  Compass,
   outcome_set:        Target,
   direction_approved: Compass,
   qa_review:          SearchCheck,
@@ -36,6 +38,8 @@ const STATUS_TONE: Record<CaseStatus, string> = {
   new:               "bg-rule-soft text-ink-muted border-rule",
   processing:        "bg-info-bg text-info border-info/30",
   documents_ready:   "bg-info-bg text-info border-info/40",
+  analyst_verified:  "bg-gold-wash text-gold-deep border-gold/40",
+  research_complete: "bg-gold-wash text-gold-deep border-gold/40",
   outcome_set:       "bg-gold-wash text-gold-deep border-gold/40",
   direction_approved:"bg-gold-wash text-gold-deep border-gold/50",
   qa_review:         "bg-warn-bg text-warn border-warn/40",

web-ui/src/components/cases/status-changer.tsx

@@ -21,9 +21,13 @@ const ALL_STATUSES: readonly CaseStatus[] = CASE_STATUSES;
 export function StatusChanger({
   caseNumber,
   currentStatus,
+  inline = false,
 }: {
   caseNumber: string;
   currentStatus?: CaseStatus;
+  /** Compact horizontal layout for the status hero (mockup 18b hero-controls);
+   *  default is the stacked rail layout. */
+  inline?: boolean;
 }) {
   // `null` = untouched → the dropdown tracks the live `currentStatus` (which
   // arrives async and changes on the 5s poll / external updates). Only an
@@ -45,8 +49,10 @@ export function StatusChanger({
   };
 
   return (
-    <div className="mt-4 border-t border-rule pt-3 space-y-2">
+    <div className={inline ? "space-y-0" : "mt-4 border-t border-rule pt-3 space-y-2"}>
+      {!inline && (
         <label className="text-[0.72rem] text-ink-muted block">שינוי סטטוס ידני</label>
+      )}
       <div className="flex items-center gap-2">
         <Select
           value={effective || "__current__"}
@@ -77,7 +83,7 @@ export function StatusChanger({
           disabled={!canSave || mutate.isPending}
           onClick={handleSave}
         >
-          {mutate.isPending ? "שומר…" : "עדכן"}
+          {mutate.isPending ? "שומר…" : inline ? "עדכן סטטוס" : "עדכן"}
         </Button>
       </div>
     </div>

web-ui/src/components/cases/status-rail.tsx

@@ -0,0 +1,195 @@
+"use client";
+
+import {
+  PHASES,
+  phaseIndexOf,
+  phaseOf,
+  type CaseStatus,
+} from "@/lib/api/case-status";
+import { StatusChanger } from "@/components/cases/status-changer";
+import { StatusGuide } from "@/components/cases/status-guide";
+import { AgentStatusWidget } from "@/components/cases/agent-status-widget";
+import { SyncIndicator } from "@/components/cases/sync-indicator";
+import {
+  Popover,
+  PopoverContent,
+  PopoverTrigger,
+} from "@/components/ui/popover";
+import { expectedOutcomes } from "@/lib/schemas/case";
+import { formatDate } from "@/lib/format-date";
+import { Settings2, Users } from "lucide-react";
+import type { CaseDetail } from "@/lib/api/cases";
+
+/*
+ * StatusRail — the integrated case-status strip that lives inside the parchment
+ * band (CaseHeader), so it is visible on EVERY tab (banner "A", X17 case-detail
+ * redesign — supersedes the overview-only StatusHero). One thin row of chips
+ * (current phase · expected outcome · meta · agents popover · manual changer)
+ * over the 5-phase pipeline stepper. Reuses the shared StatusChanger / StatusGuide
+ * / AgentStatusWidget so no parallel status path is created (G2).
+ */
+
+const EXPECTED_OUTCOME_LABELS: Record<string, string> = Object.fromEntries(
+  expectedOutcomes.map((o) => [o.value, o.label]),
+);
+
+/** Semantic colour per expected-outcome value (rejection/partial/full/none). */
+const OUTCOME_TONE: Record<string, string> = {
+  rejection: "bg-danger-bg text-danger",
+  partial_acceptance: "bg-warn-bg text-warn",
+  full_acceptance: "bg-success-bg text-success",
+};
+
+/** Horizontal RTL stepper over the 5 lifecycle phases (mockup A .stepper). */
+function HorizontalStepper({ status }: { status?: CaseStatus }) {
+  const currentIdx = phaseIndexOf(status);
+  return (
+    <ol className="flex items-start gap-0">
+      {PHASES.map((phase, i) => {
+        const state =
+          currentIdx === -1 ? "pending"
+          : i < currentIdx ? "done"
+          : i === currentIdx ? "now"
+          : "pending";
+
+        const isLast = i === PHASES.length - 1;
+
+        const knobTone =
+          state === "done"
+            ? "bg-success [box-shadow:0_0_0_1px_var(--color-success)]"
+            : state === "now"
+              ? "bg-gold [box-shadow:0_0_0_2px_var(--color-gold-wash),0_0_0_3px_var(--color-gold)]"
+              : "bg-rule [box-shadow:0_0_0_1px_var(--color-rule)]";
+
+        const labelTone =
+          state === "done"
+            ? "text-ink-soft"
+            : state === "now"
+              ? "text-navy font-bold"
+              : "text-ink-muted";
+
+        const segTone = state === "done" ? "bg-success" : "bg-rule";
+
+        return (
+          <li key={phase.key} className="relative flex-1 pt-6 text-center">
+            {!isLast && (
+              <span
+                className={`absolute top-[7px] h-0.5 w-full ${segTone}`}
+                style={{ insetInlineStart: "50%" }}
+                aria-hidden
+              />
+            )}
+            <span
+              className={`absolute top-[2px] h-3.5 w-3.5 rounded-full border-[2.5px] border-surface ${knobTone}`}
+              style={{ insetInlineStart: "50%", transform: "translateX(50%)" }}
+              aria-hidden
+            />
+            <span className={`block text-[0.74rem] ${labelTone}`}>
+              {phase.label}
+            </span>
+          </li>
+        );
+      })}
+    </ol>
+  );
+}
+
+const PILL =
+  "inline-flex items-center gap-1.5 rounded-full border border-rule bg-surface px-3 py-1 text-[0.72rem] font-semibold text-ink-soft hover:bg-parchment transition-colors";
+
+export function StatusRail({
+  caseNumber,
+  data,
+}: {
+  caseNumber: string;
+  data?: CaseDetail;
+}) {
+  const status = data?.status;
+  const phaseKey = phaseOf(status);
+  const phaseLabel = phaseKey
+    ? PHASES.find((p) => p.key === phaseKey)?.label
+    : null;
+  const outcome = data?.expected_outcome ?? "";
+  const outcomeLabel = outcome
+    ? EXPECTED_OUTCOME_LABELS[outcome] ?? outcome
+    : null;
+  const outcomeTone = outcome
+    ? OUTCOME_TONE[outcome] ?? "bg-warn-bg text-warn"
+    : "bg-rule-soft text-ink-muted";
+
+  return (
+    <div className="mt-3 border-t border-rule-soft pt-3">
+      {/* chip row */}
+      <div className="flex items-center gap-x-3 gap-y-2 flex-wrap mb-3">
+        <span className="text-navy text-[0.8rem] font-bold me-1">סטטוס</span>
+        {phaseLabel && (
+          <span className="rounded-full bg-warn-bg text-warn text-[0.78rem] font-bold px-3 py-1">
+            כעת: {phaseLabel}
+          </span>
+        )}
+        <span className="inline-flex items-center gap-1.5">
+          <span className="text-[0.72rem] text-ink-muted">תוצאה צפויה:</span>
+          <span
+            className={`inline-block rounded-full text-[0.74rem] font-semibold px-2.5 py-0.5 ${outcomeTone}`}
+          >
+            {outcomeLabel ?? "טרם נקבעה"}
+          </span>
+        </span>
+
+        <span className="flex-1" />
+
+        <span className="flex items-center gap-3 text-[0.72rem] text-ink-muted">
+          <span>
+            דיון{" "}
+            <b className="text-ink-soft font-semibold tabular-nums">
+              {formatDate(data?.hearing_date)}
+            </b>
+          </span>
+          <span>
+            עודכן{" "}
+            <b className="text-ink-soft font-semibold tabular-nums">
+              {formatDate(data?.updated_at)}
+            </b>
+          </span>
+          <span className="inline-flex items-center gap-1">
+            סנכרון <SyncIndicator caseNumber={data?.case_number} />
+          </span>
+        </span>
+
+        {/* agents popover */}
+        <Popover>
+          <PopoverTrigger asChild>
+            <button type="button" className={PILL} title="סוכני התיק">
+              <span className="h-1.5 w-1.5 rounded-full bg-success" />
+              <Users className="w-3.5 h-3.5" />
+              סוכנים
+            </button>
+          </PopoverTrigger>
+          <PopoverContent align="end" className="w-80 p-3">
+            <AgentStatusWidget caseNumber={caseNumber} />
+          </PopoverContent>
+        </Popover>
+
+        {/* manual status changer + status guide popover */}
+        <Popover>
+          <PopoverTrigger asChild>
+            <button type="button" className={PILL} title="שינוי סטטוס ידני">
+              <Settings2 className="w-3.5 h-3.5" />
+              שינוי
+            </button>
+          </PopoverTrigger>
+          <PopoverContent align="end" className="w-72 p-3 space-y-3">
+            <div className="text-[0.72rem] font-semibold text-ink-muted">
+              שינוי סטטוס ידני
+            </div>
+            <StatusChanger caseNumber={caseNumber} currentStatus={status} inline />
+            <StatusGuide />
+          </PopoverContent>
+        </Popover>
+      </div>
+
+      {/* pipeline */}
+      <HorizontalStepper status={status} />
+    </div>
+  );
+}

web-ui/src/components/cases/sync-indicator.tsx

@@ -2,17 +2,7 @@
 
 import { useGitStatus } from "@/lib/api/cases";
 import { CheckCircle2, AlertCircle, Clock, CloudOff } from "lucide-react";
-
-function formatRelative(iso: string): string {
-  const diff = Date.now() - new Date(iso).getTime();
-  const mins = Math.floor(diff / 60_000);
-  if (mins < 1) return "עכשיו";
-  if (mins < 60) return `לפני ${mins} דק׳`;
-  const hours = Math.floor(mins / 60);
-  if (hours < 24) return `לפני ${hours} שע׳`;
-  const days = Math.floor(hours / 24);
-  return `לפני ${days} ימים`;
-}
+import { formatRelative } from "@/lib/format-date";
 
 export function SyncIndicator({ caseNumber }: { caseNumber?: string }) {
   const { data, isLoading } = useGitStatus(caseNumber);

web-ui/src/components/cases/workflow-timeline.tsx

@@ -1,6 +1,6 @@
 "use client";
 
-import { PHASES, type CaseStatus, type PhaseKey } from "@/lib/api/case-status";
+import { PHASES, phaseIndexOf, type CaseStatus, type PhaseKey } from "@/lib/api/case-status";
 import { STATUS_LABELS, STATUS_ICONS, STATUS_DESCRIPTIONS } from "@/components/cases/status-badge";
 import {
   FolderInput, ClipboardList, Brain, PenLine, CheckCircle2,
@@ -21,11 +21,6 @@ const PHASE_ICONS: Record<PhaseKey, LucideIcon> = {
   done:     CheckCircle2,
 };
 
-function phaseIndexOf(status?: CaseStatus): number {
-  if (!status) return -1;
-  return PHASES.findIndex((p) => p.statuses.includes(status));
-}
-
 export function WorkflowTimeline({ status }: { status?: CaseStatus }) {
   const currentIdx = phaseIndexOf(status);
 

web-ui/src/components/compose/chair-editor.tsx

@@ -2,6 +2,7 @@
 
 import { useEffect, useRef, useState } from "react";
 import { useSaveChairPosition } from "@/lib/api/research";
+import { formatTime } from "@/lib/format-date";
 
 /*
  * Chair-position editor for a single threshold claim or issue.
@@ -87,10 +88,7 @@ function SaveIndicator({ state }: { state: SaveState }) {
     return <span className="text-[0.72rem] text-ink-muted">⏳ שומר…</span>;
   }
   if (state.kind === "saved") {
-    const time = state.at.toLocaleTimeString("he-IL", {
-      hour: "2-digit",
-      minute: "2-digit",
-    });
+    const time = formatTime(state.at);
     return <span className="text-[0.72rem] text-success">✓ נשמר {time}</span>;
   }
   return <span className="text-[0.72rem] text-danger">⚠ {state.message}</span>;

web-ui/src/components/compose/citation-verification-panel.tsx

@@ -5,12 +5,22 @@
  * supporting precedents with the cumulative authority signal (cited_by), a verify
  * gate + chair note, and per-issue radar (unlinked digests). The writer cites only
  * verified quotes (INV-AH); the digest is never cited (INV-DIG1, radar only).
+ *
+ * Redesign (mockup 18g): summary + verification progress bar, status filters,
+ * collapsible per-argument accordion with a verify-progress chip, and compact
+ * precedent rows (clamped quote, collapsible chair note) — easier to triage.
  */
 
 import { useState } from "react";
 import { toast } from "sonner";
 import { Card, CardContent } from "@/components/ui/card";
 import { Skeleton } from "@/components/ui/skeleton";
+import {
+  Accordion,
+  AccordionContent,
+  AccordionItem,
+  AccordionTrigger,
+} from "@/components/ui/accordion";
 import {
   useCitationVerification,
   useVerifyCitation,
@@ -33,10 +43,34 @@ const RADAR_LABEL: Record<string, string> = {
   available_link: "בקורפוס — לקשר",
 };
 
+type FilterKey = "all" | "pending" | "verified" | "radar";
+const FILTERS: { key: FilterKey; label: string }[] = [
+  { key: "all", label: "הכל" },
+  { key: "pending", label: "ממתין לאימות" },
+  { key: "verified", label: "מאומת" },
+  { key: "radar", label: "עם רדאר" },
+];
+
 export function CitationVerificationPanel({ caseNumber }: { caseNumber: string }) {
   const { data, isPending, isError } = useCitationVerification(caseNumber);
   const verify = useVerifyCitation(caseNumber);
   const [notes, setNotes] = useState<Record<string, string>>({});
+  const [openNotes, setOpenNotes] = useState<Set<string>>(new Set());
+  const [expandedQuotes, setExpandedQuotes] = useState<Set<string>>(new Set());
+  const [filter, setFilter] = useState<FilterKey>("all");
+  const [onlyUnverified, setOnlyUnverified] = useState(false);
+
+  function toggleSet(
+    setter: React.Dispatch<React.SetStateAction<Set<string>>>,
+    id: string,
+  ) {
+    setter((prev) => {
+      const next = new Set(prev);
+      if (next.has(id)) next.delete(id);
+      else next.add(id);
+      return next;
+    });
+  }
 
   if (isPending) {
     return (
@@ -111,76 +145,178 @@ export function CitationVerificationPanel({ caseNumber }: { caseNumber: string }
   }
 
   const sum = data.summary;
+  const args = data.arguments;
+
+  // verification progress across all supporting precedents
+  let totalSup = 0;
+  let verifiedSup = 0;
+  for (const a of args) {
+    totalSup += a.supporting.length;
+    verifiedSup += a.supporting.filter((s) => s.verified).length;
+  }
+  const pendingSup = totalSup - verifiedSup;
+  const okPct = totalSup ? Math.round((verifiedSup / totalSup) * 100) : 0;
+  const supPct = totalSup ? Math.round((pendingSup / totalSup) * 100) : 0;
+  const noSupportCount = sum.arguments_total - sum.arguments_with_support;
+
+  const argVerified = (a: ArgumentBlock) =>
+    a.supporting.filter((s) => s.verified).length;
+  const argHasPending = (a: ArgumentBlock) =>
+    a.supporting.some((s) => !s.verified);
+
+  const shownArgs = args.filter((a) => {
+    if (filter === "pending") return argHasPending(a);
+    if (filter === "verified") return argVerified(a) > 0;
+    if (filter === "radar") return a.radar.length > 0;
+    return true;
+  });
+
+  // default-open the actionable arguments (those with unverified support)
+  const defaultOpen = args.filter(argHasPending).map((a) => a.argument_id);
 
   return (
     <div className="space-y-4">
-      {/* summary */}
-      <div className="flex flex-wrap items-center gap-x-5 gap-y-1 text-sm text-ink-soft bg-parchment border border-rule rounded-lg px-4 py-2.5">
-        <span>סוגיות עם פסיקה: <b className="text-navy tabular-nums">{sum.arguments_with_support}/{sum.arguments_total}</b></span>
-        <span>אומתו: <b className="text-success tabular-nums">{sum.verified}</b></span>
-        <span>לידֵי רדאר: <b className="text-warn tabular-nums">{sum.radar_leads}</b></span>
-        <span className="text-ink-muted text-[0.78rem] ms-auto">ה-writer מצטט רק מאומתים (INV-AH) · היומון אינו מצוטט (INV-DIG1)</span>
+      {/* summary + progress */}
+      <div className="rounded-xl border border-rule bg-surface shadow-sm px-5 py-4">
+        <div className="flex items-center gap-6 flex-wrap">
+          <Stat k="סוגיות עם פסיקה" v={`${sum.arguments_with_support}/${sum.arguments_total}`} />
+          <Stat k="אומתו" v={String(sum.verified)} tone="text-success" />
+          <Stat k="לידֵי רדאר" v={String(sum.radar_leads)} tone="text-warn" />
+          <span className="ms-auto text-[0.72rem] text-ink-muted max-w-[280px] leading-snug">
+            ה-writer מצטט רק מאומתים (INV-AH) · היומון אינו מצוטט (INV-DIG1)
+          </span>
+        </div>
+        <div className="mt-3.5">
+          <div className="h-2 rounded-full bg-rule-soft overflow-hidden flex">
+            <span className="bg-success h-full" style={{ width: `${okPct}%` }} />
+            <span className="bg-gold-soft h-full" style={{ width: `${supPct}%` }} />
+          </div>
+          <div className="flex gap-4 mt-1.5 text-[0.7rem] text-ink-muted flex-wrap">
+            <span><Dot className="bg-success" /><b className="text-ink-soft">{verifiedSup} מאומתים</b></span>
+            <span><Dot className="bg-gold-soft" /><b className="text-ink-soft">{pendingSup} ממתינים לאימות</b></span>
+            <span><Dot className="bg-rule-soft" />{noSupportCount} סוגיות ללא פסיקה תומכת</span>
+          </div>
+        </div>
       </div>
 
-      {data.arguments.map((a) => (
-        <Card key={a.argument_id} className="bg-surface border-rule shadow-sm overflow-hidden">
-          <CardContent className="p-0">
-            {/* issue header */}
-            <div className="flex items-start gap-3 px-5 py-3.5 border-b border-rule-soft bg-parchment/60">
-              <div className="min-w-0 flex-1">
-                <div className="text-navy font-bold text-[0.98rem]">{a.title}</div>
-                <div className="text-ink-muted text-xs mt-0.5">
+      {/* filters */}
+      <div className="flex items-center gap-2 flex-wrap">
+        {FILTERS.map((f) => (
+          <button
+            key={f.key}
+            type="button"
+            onClick={() => setFilter(f.key)}
+            className={`text-[0.78rem] font-semibold rounded-full px-3.5 py-1.5 border transition-colors ${
+              filter === f.key
+                ? "bg-navy text-parchment border-transparent"
+                : "bg-surface text-ink-soft border-rule hover:bg-parchment"
+            }`}
+          >
+            {f.label}
+          </button>
+        ))}
+        <label className="ms-auto flex items-center gap-2 text-[0.78rem] text-ink-soft font-medium cursor-pointer select-none">
+          רק לא-מאומתים
+          <input
+            type="checkbox"
+            checked={onlyUnverified}
+            onChange={(e) => setOnlyUnverified(e.target.checked)}
+            className="accent-gold w-4 h-4"
+          />
+        </label>
+      </div>
+
+      {/* arguments */}
+      <Accordion type="multiple" defaultValue={defaultOpen} className="space-y-2.5">
+        {shownArgs.map((a) => {
+          const v = argVerified(a);
+          const t = a.supporting.length;
+          const tone =
+            t === 0 || v === 0
+              ? "bg-rule-soft text-ink-muted"
+              : v === t
+                ? "bg-success-bg text-success"
+                : "bg-warn-bg text-warn";
+          const supShown = onlyUnverified
+            ? a.supporting.filter((s) => !s.verified)
+            : a.supporting;
+          return (
+            <AccordionItem
+              key={a.argument_id}
+              value={a.argument_id}
+              className="border border-rule rounded-xl bg-surface shadow-sm overflow-hidden data-[state=open]:border-gold-soft"
+            >
+              <AccordionTrigger className="px-4 py-3 hover:no-underline gap-3">
+                <span className="flex flex-1 items-center gap-3 min-w-0">
+                  <span className="flex flex-col items-start min-w-0 text-start">
+                    <span className="text-navy font-bold text-[0.92rem] truncate max-w-full">
+                      {a.title}
+                    </span>
+                    <span className="text-ink-muted text-[0.72rem]">
                       סוגיה{a.legal_topic ? ` · ${a.legal_topic}` : ""}
-                </div>
-              </div>
+                    </span>
+                  </span>
                   {a.priority && (
-                <span className="shrink-0 rounded-full bg-gold-wash text-gold-deep text-[0.68rem] font-semibold px-2.5 py-0.5">
+                    <span className="shrink-0 rounded-full bg-gold-wash text-gold-deep text-[0.66rem] font-semibold px-2.5 py-0.5">
                       {PRIORITY_LABEL[a.priority] ?? a.priority}
                     </span>
                   )}
+                  <span className={`shrink-0 rounded-full text-[0.7rem] font-bold px-2.5 py-0.5 tabular-nums ${tone}`}>
+                    {v}/{t} ✓
+                  </span>
+                </span>
+              </AccordionTrigger>
+              <AccordionContent className="px-4 pb-3 pt-0">
+                <div className="text-[0.72rem] font-bold text-ink-muted mb-1">
+                  פסיקה תומכת בקורפוס
                 </div>
-
-            {/* supporting precedents */}
-            <div className="px-5 py-3">
-              <div className="text-[0.72rem] font-bold text-ink-muted mb-1">פסיקה תומכת בקורפוס</div>
                 {a.supporting.length === 0 ? (
-                <p className="text-ink-muted text-sm py-2">לא נמצאה פסיקה תומכת בקורפוס לסוגיה זו.</p>
+                  <p className="text-ink-muted text-sm py-2">
+                    לא נמצאה פסיקה תומכת בקורפוס לסוגיה זו.
+                  </p>
+                ) : supShown.length === 0 ? (
+                  <p className="text-ink-muted text-sm py-2">
+                    כל הפסיקה לסוגיה זו אומתה.
+                  </p>
                 ) : (
-                <ul className="list-none p-0 m-0 divide-y divide-rule-soft">
-                  {a.supporting.map((s) => (
-                    <li key={s.case_law_id} className="py-3 first:pt-1">
+                  <ul className="list-none p-0 m-0">
+                    {supShown.map((s) => {
+                      const noteOpen =
+                        openNotes.has(s.case_law_id) || !!s.chair_note;
+                      const quoteExpanded = expandedQuotes.has(s.case_law_id);
+                      return (
+                        <li
+                          key={s.case_law_id}
+                          className="py-3 border-t border-rule-soft first:border-t-0"
+                        >
                           <div className="flex items-center gap-2 flex-wrap">
                             <a
                               href={`/precedents/${s.case_law_id}`}
-                          className="font-bold text-gold-deep text-[0.86rem] hover:text-navy"
+                              className="font-bold text-gold-deep text-[0.85rem] hover:text-navy"
                             >
                               {s.case_number || s.case_name}
                             </a>
                             {s.cited_by.positive > 0 && (
-                          <span className="rounded-full bg-success-bg text-success text-[0.68rem] font-semibold px-2 py-0.5">
+                              <span className="rounded-full bg-success-bg text-success text-[0.66rem] font-semibold px-2 py-0.5">
                                 אומץ ×{s.cited_by.positive}
                               </span>
                             )}
                             {s.cited_by.negative > 0 && (
-                          <span className="rounded-full bg-danger-bg text-danger text-[0.68rem] font-semibold px-2 py-0.5">
+                              <span className="rounded-full bg-danger-bg text-danger text-[0.66rem] font-semibold px-2 py-0.5">
                                 אובחן ×{s.cited_by.negative}
                               </span>
                             )}
-                      </div>
-                      {s.quote && (
-                        <blockquote className="border-s-[3px] border-gold bg-gold-wash text-ink-soft text-sm leading-7 rounded-e-md px-3.5 py-2 my-2 mx-0">
-                          {s.quote}
-                        </blockquote>
-                      )}
-                      <div className="flex items-center gap-2 flex-wrap">
+                            <span className="ms-auto flex items-center gap-1.5">
                               {s.verified ? (
                                 <>
-                            <span className="text-success text-xs font-semibold">✓ אומת ע״י היו״ר</span>
+                                  <span className="text-success text-[0.72rem] font-semibold">
+                                    ✓ אומת ע״י היו״ר
+                                  </span>
                                   <button
                                     type="button"
                                     disabled={verify.isPending}
                                     onClick={() => runVerify(a, s, false)}
-                              className="text-xs text-ink-muted hover:text-danger border border-rule rounded-md px-2.5 py-1 disabled:opacity-40"
+                                    className="text-[0.72rem] text-ink-muted hover:text-danger border border-rule rounded-md px-2.5 py-1 disabled:opacity-40"
                                   >
                                     בטל אימות
                                   </button>
@@ -191,7 +327,7 @@ export function CitationVerificationPanel({ caseNumber }: { caseNumber: string }
                                     type="button"
                                     disabled={verify.isPending}
                                     onClick={() => runVerify(a, s, true)}
-                              className="text-xs font-semibold text-success bg-success-bg border border-success/30 rounded-md px-3 py-1 hover:brightness-95 disabled:opacity-40"
+                                    className="text-[0.72rem] font-semibold text-success bg-success-bg border border-success/30 rounded-md px-3 py-1 hover:brightness-95 disabled:opacity-40"
                                   >
                                     ✓ מאמת
                                   </button>
@@ -199,49 +335,84 @@ export function CitationVerificationPanel({ caseNumber }: { caseNumber: string }
                                     type="button"
                                     disabled={verify.isPending}
                                     onClick={() => runVerify(a, s, false)}
-                              className="text-xs font-semibold text-danger border border-rule rounded-md px-3 py-1 hover:bg-danger-bg disabled:opacity-40"
+                                    className="text-[0.72rem] font-semibold text-danger border border-rule rounded-md px-3 py-1 hover:bg-danger-bg disabled:opacity-40"
                                   >
                                     ✗ לא רלוונטי
                                   </button>
                                 </>
                               )}
-                        {s.cited_by.negative > 0 && !s.verified && (
-                          <span className="text-warn text-[0.72rem] ms-1">⚠ אובחן — בדוק הקשר</span>
-                        )}
+                            </span>
                           </div>
-                      {/* chair note */}
-                      <div className="mt-2 flex items-start gap-2">
-                        <label className="text-[0.72rem] font-bold text-gold-deep whitespace-nowrap pt-1.5">
+                          {s.quote && (
+                            <blockquote
+                              onClick={() => toggleSet(setExpandedQuotes, s.case_law_id)}
+                              title="לחצי להרחבה/כיווץ"
+                              className={`mt-2 cursor-pointer border-s-[3px] border-gold bg-gold-wash text-ink-soft text-[0.82rem] leading-6 rounded-e-md px-3 py-1.5 ${quoteExpanded ? "" : "line-clamp-2"}`}
+                            >
+                              {s.quote}
+                            </blockquote>
+                          )}
+                          {s.cited_by.negative > 0 && !s.verified && (
+                            <div className="text-warn text-[0.72rem] mt-1.5">⚠ אובחן — בדוק הקשר</div>
+                          )}
+                          {noteOpen ? (
+                            <div className="mt-2 flex items-center gap-2">
+                              <label className="text-[0.72rem] font-bold text-gold-deep whitespace-nowrap">
                                 הערת יו״ר:
                               </label>
                               <input
                                 type="text"
                                 defaultValue={s.chair_note}
                                 placeholder="הוסף הערה לציטוט — מדוע תומך / הסתייגות / איך לשלב…"
-                          onChange={(e) => setNotes((n) => ({ ...n, [s.case_law_id]: e.target.value }))}
+                                onChange={(e) =>
+                                  setNotes((n) => ({ ...n, [s.case_law_id]: e.target.value }))
+                                }
                                 onBlur={() => saveNote(a, s)}
                                 className="flex-1 text-[0.8rem] text-ink-soft bg-parchment border border-dashed border-rule rounded-md px-2.5 py-1.5 focus:outline-none focus:border-gold"
                               />
                             </div>
+                          ) : (
+                            <button
+                              type="button"
+                              onClick={() => toggleSet(setOpenNotes, s.case_law_id)}
+                              className="mt-2 text-[0.72rem] text-gold-deep font-semibold hover:underline"
+                            >
+                              ＋ הוסף הערת יו״ר
+                            </button>
+                          )}
                         </li>
-                  ))}
+                      );
+                    })}
                   </ul>
                 )}
-            </div>
 
-            {/* radar — unlinked digests for this issue (INV-DIG1: pointer, not citation) */}
                 <RadarStrip radar={a.radar} />
-          </CardContent>
-        </Card>
-      ))}
+              </AccordionContent>
+            </AccordionItem>
+          );
+        })}
+      </Accordion>
     </div>
   );
 }
 
+function Stat({ k, v, tone }: { k: string; v: string; tone?: string }) {
+  return (
+    <span className="flex flex-col">
+      <span className="text-[0.72rem] text-ink-muted">{k}</span>
+      <span className={`text-lg font-bold tabular-nums mt-0.5 ${tone ?? "text-navy"}`}>{v}</span>
+    </span>
+  );
+}
+
+function Dot({ className }: { className: string }) {
+  return <span className={`inline-block w-2 h-2 rounded-sm me-1 align-middle ${className}`} />;
+}
+
 function RadarStrip({ radar }: { radar: RadarLead[] }) {
   if (!radar.length) return null;
   return (
-    <div className="mx-5 mb-4 rounded-lg border border-dashed border-gold bg-parchment px-4 py-2.5">
+    <div className="mt-3 rounded-lg border border-dashed border-gold bg-parchment px-4 py-2.5">
       <div className="text-[0.72rem] font-bold text-gold-deep mb-1">
         📡 רדאר — פסיקה שאין לנו בקורפוס (מצביע, לא מצוטט)
       </div>

web-ui/src/components/digests/digest-card.tsx

@@ -5,15 +5,7 @@ import type { ReactNode } from "react";
 import { Badge } from "@/components/ui/badge";
 import { practiceAreaLabel } from "@/components/precedents/practice-area";
 import type { Digest } from "@/lib/api/digests";
-
-function formatDate(iso: string | null) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL");
-  } catch {
-    return iso;
-  }
-}
+import { formatDateShort as formatDate } from "@/lib/format-date";
 
 /**
  * Presentational card for a single digest ("כל יום" radar entry).

web-ui/src/components/graph/graph-filter-panel.tsx

@@ -26,6 +26,7 @@ import {
 } from "@/components/ui/select";
 import type { GraphFacets } from "@/lib/api/graph";
 import type { ColorBy, SizeBy } from "@/components/graph/graph-canvas";
+import { getIsraelYear } from "@/lib/format-date";
 
 export type GraphControls = {
   practiceArea: string;
@@ -51,7 +52,7 @@ const ALL = "__all__";
 // Floor covers the oldest dated precedent in the corpus (currently ע"א 725/81 →
 // 1982); ceiling tracks the current year so the range never ages out.
 const YEAR_FLOOR = 1980;
-const CURRENT_YEAR = new Date().getFullYear();
+const CURRENT_YEAR = getIsraelYear(new Date()) ?? new Date().getFullYear();
 const YEARS = Array.from(
   { length: CURRENT_YEAR - YEAR_FLOOR + 1 },
   (_, i) => CURRENT_YEAR - i,

web-ui/src/components/header-context.ts

@@ -3,9 +3,8 @@
  * header. Reflects the current section so the user always sees where they
  * are without scanning the nav row.
  *
- * Special-cases case routes (`/cases/{caseNumber}` and `/compose`) so the
- * subtitle includes the case number — the most useful piece of context
- * during decision drafting.
+ * Special-cases case routes (`/cases/{caseNumber}`) so the subtitle includes
+ * the case number — the most useful piece of context during decision drafting.
  */
 export function headerSubtitle(pathname: string): string {
   if (pathname === "/") return "בית";
@@ -13,9 +12,8 @@ export function headerSubtitle(pathname: string): string {
   if (pathname.startsWith("/cases/")) {
     const [, , slug] = pathname.split("/");
     if (!slug || slug === "new") return "תיק חדש";
-    const isCompose = pathname.includes("/compose");
     const decoded = decodeURIComponent(slug);
-    return isCompose ? `ערר ${decoded} · ניסוח` : `ערר ${decoded}`;
+    return `ערר ${decoded}`;
   }
 
   if (pathname.startsWith("/archive"))     return "ארכיון";

web-ui/src/components/missing-precedents/missing-precedents-table.tsx

@@ -20,15 +20,7 @@ import {
   type MissingPrecedentStatus,
 } from "@/lib/api/missing-precedents";
 import { MissingPrecedentDetailDrawer } from "./missing-precedent-detail-drawer";
-
-function formatDate(iso: string | null) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL");
-  } catch {
-    return iso;
-  }
-}
+import { formatDateShort as formatDate } from "@/lib/format-date";
 
 /** Status chip — mockup 09 tones (open=warn, uploaded=info, closed=success,
  *  irrelevant=muted). Pill-shaped, whitespace-nowrap. */

web-ui/src/components/operations/system-health-section.tsx

@@ -14,6 +14,7 @@ import { Card, CardContent } from "@/components/ui/card";
 import { Badge } from "@/components/ui/badge";
 import { Skeleton } from "@/components/ui/skeleton";
 import { useDiagnostics, type DiagDoc } from "@/lib/api/system";
+import { formatDateShort, formatRelative } from "@/lib/format-date";
 
 const TABLE_LABELS: Record<string, string> = {
   cases: "תיקים",
@@ -25,16 +26,10 @@ const TABLE_LABELS: Record<string, string> = {
 
 function formatRelativeTime(iso: string | null) {
   if (!iso) return "—";
-  const then = new Date(iso);
-  const diffMs = Date.now() - then.getTime();
-  const min = Math.floor(diffMs / 60000);
-  if (min < 1) return "עכשיו";
-  if (min < 60) return `לפני ${min} דקות`;
-  const hr = Math.floor(min / 60);
-  if (hr < 24) return `לפני ${hr} שעות`;
-  const days = Math.floor(hr / 24);
-  if (days < 30) return `לפני ${days} ימים`;
-  return then.toLocaleDateString("he-IL");
+  const min = Math.floor((Date.now() - new Date(iso).getTime()) / 60000);
+  // Past 30 days, fall back to an absolute Israel-time date.
+  if (min >= 30 * 24 * 60) return formatDateShort(iso);
+  return formatRelative(iso, { units: "long" });
 }
 
 function DocRow({ doc, tone }: { doc: DiagDoc; tone: "danger" | "warn" }) {

web-ui/src/components/precedents/halacha-review-panel.tsx

@@ -15,6 +15,7 @@ import {
   type Halacha, type CanonicalInstance,
 } from "@/lib/api/precedent-library";
 import { AuthorityBadge, ruleTypeLabel } from "./halacha-meta";
+import { formatDateShort as formatDate } from "@/lib/format-date";
 
 /** #81 strict-rubric flags — why an item was held back from auto-approval. */
 const QUALITY_FLAG_LABELS: Record<string, string> = {
@@ -28,15 +29,6 @@ const QUALITY_FLAG_LABELS: Record<string, string> = {
   nevo_preamble_leak: "דליפת רציו נבו",
 };
 
-function formatDate(iso: string | null | undefined) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL");
-  } catch {
-    return iso;
-  }
-}
-
 /* Strip Unicode bidi marks that render as zero-width but shift visual position. */
 function cleanCitation(s: string | null | undefined): string {
   if (!s) return "—";

web-ui/src/components/precedents/library-list-panel.tsx

@@ -31,15 +31,7 @@ import {
   FormattedCitation,
   CitationCopyButton,
 } from "./formatted-citation";
-
-function formatDate(iso: string | null) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL");
-  } catch {
-    return iso;
-  }
-}
+import { formatDateShort as formatDate } from "@/lib/format-date";
 
 function cleanCitation(s: string | null | undefined): string {
   if (!s) return "—";

web-ui/src/components/precedents/library-search-panel.tsx

@@ -14,15 +14,7 @@ import {
 } from "@/lib/api/precedent-library";
 import { PRACTICE_AREAS, PRECEDENT_LEVELS } from "./practice-area";
 import { AuthorityBadge } from "./halacha-meta";
-
-function formatDate(iso: string | null) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL");
-  } catch {
-    return iso;
-  }
-}
+import { formatDateShort as formatDate } from "@/lib/format-date";
 
 /** Score chip — boxed, gold-deep, tabular (mockup 07 `.score`). Sits at the
  *  end of the meta row via `ms-auto`. White fill on gold-wash halacha cards. */

web-ui/src/components/precedents/precedent-upload-sheet.tsx

@@ -23,6 +23,7 @@ import {
 } from "@/lib/api/precedent-library";
 import { ApiError } from "@/lib/api/client";
 import { useProgress } from "@/lib/api/documents";
+import { formatDateShort } from "@/lib/format-date";
 import {
   PRACTICE_AREAS, PRECEDENT_LEVELS, SOURCE_TYPES,
 } from "./practice-area";
@@ -192,7 +193,7 @@ export function PrecedentUploadSheet({ open, onOpenChange }: Props) {
                   {conflict.citation}
                   {conflict.case_name && ` — ${conflict.case_name}`}
                   {conflict.court && ` · ${conflict.court}`}
-                  {conflict.date && ` · ${new Date(conflict.date).toLocaleDateString("he-IL")}`}
+                  {conflict.date && ` · ${formatDateShort(conflict.date)}`}
                 </p>
                 {conflict.halacha_extraction_status && (
                   <p className="text-danger/60 text-[0.72rem]">

web-ui/src/components/training/chat-panel.tsx

@@ -41,6 +41,7 @@ import {
   useDeleteChat,
   type ChatMessage,
 } from "@/lib/api/training";
+import { formatDateShort } from "@/lib/format-date";
 import { useQueryClient } from "@tanstack/react-query";
 
 export function ChatPanel() {
@@ -200,7 +201,7 @@ function ConversationsSidebar({
                       <span className="tabular-nums">{c.message_count}</span>
                       <MessageSquare className="w-3 h-3" />
                       <span className="grow text-end">
-                        {new Date(c.last_message_at).toLocaleDateString("he-IL")}
+                        {formatDateShort(c.last_message_at)}
                       </span>
                       <button
                         onClick={(e) => { e.stopPropagation(); onDelete(c.id); }}

web-ui/src/components/training/corpus-panel.tsx

@@ -11,6 +11,7 @@ import { Badge } from "@/components/ui/badge";
 import { Skeleton } from "@/components/ui/skeleton";
 import { useCorpus, useDeleteCorpusEntry, type CorpusDecision } from "@/lib/api/training";
 import { CorpusDetailDrawer } from "./corpus-detail-drawer";
+import { formatDateShort as formatDate } from "@/lib/format-date";
 
 /*
  * Corpus tab: table of all decisions currently in the style corpus.
@@ -30,15 +31,6 @@ function formatChars(n: number) {
   return `${(n / 1000).toFixed(1)}K`;
 }
 
-function formatDate(iso: string) {
-  if (!iso) return "—";
-  try {
-    return new Date(iso).toLocaleDateString("he-IL");
-  } catch {
-    return iso;
-  }
-}
-
 function Row({
   item, onOpen,
 }: { item: CorpusDecision; onOpen: () => void }) {

web-ui/src/components/training/curator-portrait-panel.tsx

@@ -32,6 +32,7 @@ import { Skeleton } from "@/components/ui/skeleton";
 import { ScrollArea } from "@/components/ui/scroll-area";
 import { Tabs, TabsContent, TabsList, TabsTrigger } from "@/components/ui/tabs";
 import { Markdown } from "@/components/ui/markdown";
+import { formatDateShort, formatDateTimeFull } from "@/lib/format-date";
 import {
   useCuratorPrompt,
   useCuratorStats,
@@ -256,7 +257,7 @@ function RecentFindings() {
                 )}
                 <span className="grow text-ink-muted text-end">
                   <Clock className="w-3 h-3 inline me-1" />
-                  {new Date(f.created_at).toLocaleDateString("he-IL")}
+                  {formatDateShort(f.created_at)}
                 </span>
               </div>
               <p className="text-sm text-ink leading-relaxed">{f.lesson_text}</p>
@@ -291,7 +292,7 @@ function CuratorPromptCard() {
             <h3 className="text-navy font-semibold">{data.filename}</h3>
             <p className="text-[0.72rem] text-ink-muted">
               {data.bytes.toLocaleString("he-IL")} בייטים ·
-              עודכן: {new Date(data.last_modified * 1000).toLocaleString("he-IL")}
+              עודכן: {formatDateTimeFull(data.last_modified * 1000)}
             </p>
           </div>
           <Button asChild variant="outline" size="sm">

web-ui/src/components/training/learning-panel.tsx

@@ -9,6 +9,7 @@ import {
   useReconciliationLedger, useStyleDistance, usePairDetail, usePromoteLearning,
   type DraftFinalPair,
 } from "@/lib/api/learning";
+import { formatDateShort as fmtDate } from "@/lib/format-date";
 
 /**
  * /training "למידה" tab (T6/T13) — the reconciliation ledger (INV-LRN4): every
@@ -28,11 +29,6 @@ const STATUS_CLASS: Record<string, string> = {
   lessons_folded: "bg-emerald-50 text-emerald-800 border-emerald-300/60",
 };
 
-function fmtDate(iso: string | null) {
-  if (!iso) return "—";
-  try { return new Date(iso).toLocaleDateString("he-IL"); } catch { return iso; }
-}
-
 function StyleDistanceDetail({ caseNumber }: { caseNumber: string }) {
   const { data, isPending, error } = useStyleDistance(caseNumber);
   if (isPending) return <Loader2 className="w-4 h-4 animate-spin text-ink-muted" />;

web-ui/src/components/training/lessons-tab.tsx

@@ -37,6 +37,7 @@ import {
   usePatchLesson,
   type DecisionLesson,
 } from "@/lib/api/training";
+import { formatDateShort } from "@/lib/format-date";
 
 const CATEGORIES = [
   { value: "general", label: "כללי" },
@@ -224,7 +225,7 @@ function LessonItem({
             {reviewBadge.label}
           </Badge>
           <span className="grow text-ink-muted tabular-nums">
-            {new Date(lesson.created_at).toLocaleDateString("he-IL")}
+            {formatDateShort(lesson.created_at)}
           </span>
         </div>
 

web-ui/src/lib/api/agents.ts

@@ -26,6 +26,8 @@ export type PaperclipIssue = {
   completed_at: string | null;
   created_at: string | null;
   company_id: string;
+  /** Null for top-level issues (the CEO "start process" issue); set for sub-issues. */
+  parent_id: string | null;
 };
 
 export type PaperclipComment = {
@@ -151,8 +153,14 @@ export function useAgentActivity(caseNumber: string | undefined) {
 export function useSendComment(caseNumber: string | undefined) {
   const qc = useQueryClient();
   return useMutation({
-    mutationFn: (vars: { body: string; issue_id?: string }) =>
-      apiRequest<{ comment_id: string; issue_id: string; issue_identifier: string }>(
+    mutationFn: (vars: { body: string; issue_id?: string; new_run?: boolean }) =>
+      apiRequest<{
+        comment_id: string | null;
+        issue_id: string;
+        issue_identifier: string;
+        issue_status?: string;
+        new_run?: boolean;
+      }>(
         `/api/cases/${caseNumber}/agents/comment`,
         { method: "POST", body: vars },
       ),
@@ -180,6 +188,45 @@ export function useSubmitInteraction(caseNumber: string | undefined) {
   });
 }
 
+/** Dismiss a pending interaction WITHOUT waking the issue assignee — used for
+ * stale/duplicate questions. Backend marks it `cancelled` directly (the
+ * `wake_assignee` continuation policy means a normal resolve would wake the
+ * agent; dismiss must not). */
+export function useDismissInteraction(caseNumber: string | undefined) {
+  const qc = useQueryClient();
+  return useMutation({
+    mutationFn: (vars: { issue_id: string; interaction_id: string }) =>
+      apiRequest<{ ok: boolean; cancelled: boolean }>(
+        `/api/cases/${caseNumber}/agents/interaction-dismiss`,
+        { method: "POST", body: vars },
+      ),
+    onSuccess: () => {
+      if (caseNumber) {
+        qc.invalidateQueries({ queryKey: agentKeys.activity(caseNumber) });
+      }
+    },
+  });
+}
+
+/** Manually close a board issue to done/cancelled — the chair tidying the agents
+ * board (e.g. superseded runs left in in_review). Backend does a loop-safe
+ * direct close and issues no wakeup; only the two terminal statuses are allowed. */
+export function useSetIssueStatus(caseNumber: string | undefined) {
+  const qc = useQueryClient();
+  return useMutation({
+    mutationFn: (vars: { issue_id: string; status: "done" | "cancelled" }) =>
+      apiRequest<{ ok: boolean; id: string; identifier: string; status: string }>(
+        `/api/cases/${caseNumber}/agents/issue-status`,
+        { method: "POST", body: vars },
+      ),
+    onSuccess: () => {
+      if (caseNumber) {
+        qc.invalidateQueries({ queryKey: agentKeys.activity(caseNumber) });
+      }
+    },
+  });
+}
+
 export type AgentResetResult = {
   ok: boolean;
   reassigned_issues: { id: string; identifier: string }[];

web-ui/src/lib/api/case-files.ts

@@ -0,0 +1,66 @@
+/**
+ * Case files API — browse & read every content subfolder of a case on disk.
+ *
+ * Surfaces the agent-produced work files (drafts, OCR text, research, originals…)
+ * that previously lived only on the server filesystem. Backed by the pre-existing
+ * `/local-files` endpoints, generalized to all subfolders (path-safe on the
+ * server via `_resolve_case_file`).
+ */
+
+import { useQuery } from "@tanstack/react-query";
+import { apiRequest } from "./client";
+
+export type CaseFile = {
+  filename: string;
+  size: number;
+  /** Unix epoch seconds (float) of last modification. */
+  modified_at: number;
+};
+
+export type CaseFileFolder = {
+  /** Opaque folder key for the read endpoint (e.g. "documents__research"). */
+  key: string;
+  /** Leaf folder name (e.g. "research"). */
+  name: string;
+  /** Case-relative path (e.g. "documents/research"). */
+  path: string;
+  files: CaseFile[];
+};
+
+export type CaseFilesResponse = { folders: CaseFileFolder[] };
+
+export function useCaseFiles(caseNumber: string | undefined) {
+  return useQuery({
+    queryKey: ["case-files", caseNumber ?? ""],
+    queryFn: () =>
+      apiRequest<CaseFilesResponse>(`/api/cases/${caseNumber}/local-files`),
+    enabled: !!caseNumber,
+  });
+}
+
+/** Direct URL to a single file — use as an <a href> for download/open. */
+export function caseFileUrl(
+  caseNumber: string,
+  folderKey: string,
+  filename: string,
+): string {
+  return `/api/cases/${encodeURIComponent(caseNumber)}/local-files/${encodeURIComponent(folderKey)}/${encodeURIComponent(filename)}`;
+}
+
+/** Fetch a text file's contents (md/txt/json) for the inline viewer. */
+export function useCaseFileText(
+  caseNumber: string | undefined,
+  folderKey: string | undefined,
+  filename: string | undefined,
+) {
+  return useQuery({
+    queryKey: ["case-file-text", caseNumber ?? "", folderKey ?? "", filename ?? ""],
+    queryFn: async () => {
+      const res = await fetch(caseFileUrl(caseNumber!, folderKey!, filename!));
+      if (!res.ok) throw new Error(`שגיאה בטעינת הקובץ (${res.status})`);
+      return res.text();
+    },
+    enabled: !!caseNumber && !!folderKey && !!filename,
+    staleTime: 30_000,
+  });
+}

web-ui/src/lib/api/case-status.ts

@@ -1,16 +1,19 @@
 /**
- * Single source of truth for the case-status lifecycle (UI-B1 / G2).
+ * Frontend mirror of the canonical case-status lifecycle.
  *
- * The 17-status manual menu was trimmed to the **10 core statuses** that the
- * pipeline actually sets or that gate real logic. Decorative mid-stage markers
- * (uploading, analyst_verified, research_complete, brainstorming,
- * analysis_enriched, ready_for_writing, drafting) plus the legacy/dead
- * `in_progress` and `qa_failed` were removed — no pipeline code ever set them.
+ * The ONE source of truth is the backend status model
+ * (`mcp-server/src/legal_mcp/case_status_model.py`), exposed at
+ * `GET /api/status-model`; this file mirrors it (order / labels / phases) so the
+ * UI keeps compile-time `CaseStatus` typing. Keep the two in sync — the backend
+ * registry is authoritative (it also declares per-status `on_enter` actions).
  *
- * Every status consumer (badge, changer, timeline, guide, donut, KPI cards,
- * compose chip) imports the list / labels / phases from here instead of
- * re-declaring its own array. Keep this file in sync with the backend
- * STATUS_ORDER in `mcp-server/src/legal_mcp/tools/cases.py`.
+ * The analyst/research intermediate states (`analyst_verified`,
+ * `research_complete`) are first-class canonical statuses — the agents set them,
+ * so the chip, stepper and manual changer all agree instead of the status
+ * falling between the canonical set and a legacy bucket.
+ *
+ * Every status consumer (badge, changer, timeline, guide, donut, KPI cards)
+ * imports the list / labels / phases from here instead of re-declaring its own.
  */
 
 /** Ordered lifecycle — also the order shown in the manual status dropdown. */
@@ -18,6 +21,8 @@ export const CASE_STATUSES = [
   "new",
   "processing",
   "documents_ready",
+  "analyst_verified",
+  "research_complete",
   "outcome_set",
   "direction_approved",
   "qa_review",
@@ -35,21 +40,49 @@ export type PhaseKey = "intake" | "prep" | "thinking" | "writing" | "done";
 export const PHASES: { key: PhaseKey; label: string; statuses: CaseStatus[] }[] = [
   { key: "intake",   label: "קליטה ועיבוד", statuses: ["new", "processing"] },
   { key: "prep",     label: "הכנת תיק",      statuses: ["documents_ready"] },
-  { key: "thinking", label: "ניתוח וכיוון",  statuses: ["outcome_set", "direction_approved"] },
+  { key: "thinking", label: "ניתוח וכיוון",  statuses: ["analyst_verified", "research_complete", "outcome_set", "direction_approved"] },
   { key: "writing",  label: "כתיבת טיוטה",   statuses: ["qa_review", "drafted"] },
   { key: "done",     label: "סגירה",         statuses: ["exported", "reviewed", "final"] },
 ];
 
-/** Which phase a status belongs to (undefined for unknown/legacy values). */
+/**
+ * Truly-legacy statuses that are NOT in the canonical set but might still arrive
+ * from an old client / un-migrated row — mapped to a phase for display only, so
+ * the stepper never goes blank. (The analyst/research intermediates are NO
+ * longer here — they are canonical statuses in CASE_STATUSES above.)
+ */
+const LEGACY_STATUS_PHASE: Record<string, PhaseKey> = {
+  analysis_enriched: "thinking",
+  ready_for_writing: "writing",
+  drafting: "writing",
+  in_progress: "thinking",
+};
+
+/** Which phase a status belongs to (undefined only for truly unknown values). */
 export function phaseOf(status?: CaseStatus | string): PhaseKey | undefined {
   if (!status) return undefined;
-  return PHASES.find((p) => (p.statuses as string[]).includes(status))?.key;
+  return (
+    PHASES.find((p) => (p.statuses as string[]).includes(status))?.key ??
+    LEGACY_STATUS_PHASE[status]
+  );
+}
+
+/**
+ * Index of a status within the 5-phase pipeline (−1 only for unknown values).
+ * Single source of truth for the stepper — handles legacy statuses via phaseOf,
+ * so a case on `analyst_verified` lands on "ניתוח וכיוון" instead of blank.
+ */
+export function phaseIndexOf(status?: CaseStatus | string): number {
+  const key = phaseOf(status);
+  return key ? PHASES.findIndex((p) => p.key === key) : -1;
 }
 
 export const STATUS_LABELS: Record<CaseStatus, string> = {
   new:                "חדש",
   processing:         "בעיבוד",
   documents_ready:    "מסמכים מוכנים",
+  analyst_verified:   "ניתוח אומת",
+  research_complete:  "מחקר הושלם",
   outcome_set:        "תוצאה נקבעה",
   direction_approved: "כיוון אושר",
   qa_review:          "בדיקת איכות",
@@ -67,8 +100,6 @@ export const STATUS_LABELS: Record<CaseStatus, string> = {
 const LEGACY_STATUS_LABELS: Record<string, string> = {
   in_progress:       "בעבודה",
   uploading:         "מעלה",
-  analyst_verified:  "ניתוח אומת",
-  research_complete: "מחקר הושלם",
   brainstorming:     "סיעור מוחות",
   analysis_enriched: "ניתוח הועמק",
   ready_for_writing: "מוכן לכתיבה",
@@ -94,6 +125,8 @@ export const STATUS_DESCRIPTIONS: Record<CaseStatus, string> = {
   new:                "התיק נוצר וממתין להעלאת מסמכים",
   processing:         "המערכת מעבדת ומנתחת את המסמכים",
   documents_ready:    "כל המסמכים עובדו ומוכנים לעבודה",
+  analyst_verified:   "המנתח סיים ואימת את הניתוח — ממתין להכרעת תוצאה",
+  research_complete:  "חקר התקדימים הושלם (מסלול נפרד מהמנתח)",
   outcome_set:        "נקבעה תוצאה צפויה לערר",
   direction_approved: "כיוון ההחלטה אושר — בהעמקת ניתוח וכתיבה",
   qa_review:          "הטיוטה בבדיקת איכות אוטומטית",

web-ui/src/lib/api/citation-verification.ts

@@ -84,7 +84,13 @@ export function useCitationVerification(caseNumber: string | undefined) {
         { signal },
       ),
     enabled: Boolean(caseNumber),
-    staleTime: 10_000,
+    // The view is an expensive synthesized build (per-argument Voyage embed +
+    // vector search on the backend). Keep it fresh for 5 min and do NOT refetch
+    // on remount — Radix Tabs unmounts inactive TabsContent, so every tab switch
+    // back used to trigger a full rebuild. Verify mutations invalidate the query,
+    // so edits still pull fresh data.
+    staleTime: 5 * 60_000,
+    refetchOnMount: false,
   });
 }
 

web-ui/src/lib/api/documents.ts

@@ -92,6 +92,8 @@ export function useUploadDocument(caseNumber: string) {
 export type DocumentPatch = {
   doc_type?: string;
   appraiser_side?: string; // "" clears; "committee" | "appellant" | "deciding" sets
+  is_post_hearing?: boolean; // true = received after the hearing (→ group 4); false clears
+  protocol_scope?: string; // "" clears; "appeal" | "lower" (protocol docs only)
 };
 
 export type PatchDocumentResponse = {

web-ui/src/lib/api/generate.ts

@@ -0,0 +1,148 @@
+/**
+ * Document-generation domain hooks (TaskMaster #214).
+ *
+ * The "הפקת מסמכים" card in the case "טיוטות והערות" tab fires two deterministic,
+ * host-side generation actions through the CEO (structured-action wakeup):
+ *   • party_claims_summary → an executive summary of the parties' claims
+ *   • interim_draft        → a partial draft (blocks ה–ט, up to the discussion)
+ *
+ * Generation runs on the host (claude_session → claude -p), NOT in the container,
+ * so the trigger endpoints return 202-accepted and the UI POLLS for completion:
+ *   • summary → GET /api/cases/{n}/research/party-claims-summary (200 = ready)
+ *   • interim draft → the exports list grows a "טיוטה-טענות_הצדדים_{N}.docx" file
+ *     (polled via useExports — see exports.ts).
+ *
+ * The full-decision export ("הפק טיוטת החלטה מלאה") is NOT here — it reuses the
+ * existing in-container useExportDocx (exports.ts), no LLM, no wakeup.
+ */
+
+import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
+import { ApiError, apiRequest } from "./client";
+
+/** Shape returned by the structured-action wakeup endpoints (#214). */
+export type GenerateTriggerResult = {
+  status: "ok" | "skipped" | "error";
+  action?: string;
+  issue_id?: string;
+  ceo_id?: string;
+  reason?: string;
+  error?: string;
+};
+
+/** The two host-side background generation actions (poll target of the status chip). */
+export type GenerateAction = "party_claims_summary" | "interim_draft";
+
+export const generateKeys = {
+  all: ["generate"] as const,
+  partyClaimsSummary: (caseNumber: string) =>
+    [...generateKeys.all, "party-claims-summary", caseNumber] as const,
+  status: (caseNumber: string, action: GenerateAction) =>
+    [...generateKeys.all, "status", action, caseNumber] as const,
+};
+
+/**
+ * Server-derived background-run status (#227 ג). Reconstructed on the backend
+ * from the CEO child issue + its heartbeat_run + the output artifact — so it is
+ * correct after navigating away and back; it never depends on the browser having
+ * stayed on the page. `started_at`/`finished_at` are server ISO-8601 UTC stamps,
+ * so the client renders a correct elapsed time on return.
+ */
+export type GenerateState = "idle" | "queued" | "running" | "done" | "failed";
+
+export type GenerateStatus = {
+  case_number: string;
+  action: GenerateAction;
+  state: GenerateState;
+  started_at: string | null;
+  finished_at: string | null;
+  output_ready: boolean;
+  detail: string;
+  issue_id: string | null;
+};
+
+export function useGenerateStatus(
+  caseNumber: string | undefined,
+  action: GenerateAction,
+) {
+  return useQuery<GenerateStatus>({
+    queryKey: generateKeys.status(caseNumber ?? "", action),
+    queryFn: ({ signal }) =>
+      apiRequest<GenerateStatus>(
+        `/api/cases/${caseNumber}/generate/${action}/status`,
+        { signal },
+      ),
+    enabled: Boolean(caseNumber),
+    // Poll only while a run is in flight; a terminal state stops the interval.
+    refetchInterval: (query) => {
+      const s = query.state.data?.state;
+      return s === "running" || s === "queued" ? 5_000 : false;
+    },
+    staleTime: 2_000,
+  });
+}
+
+/* ── Read: party-claims executive summary (poll target for button 1) ──
+   404 (not generated yet) is a normal "not ready" state, not an error — we
+   coerce it to null so the panel renders the empty/idle state cleanly. */
+export type PartyClaimsSummary = { markdown: string } | null;
+
+export function usePartyClaimsSummary(caseNumber: string | undefined) {
+  return useQuery<PartyClaimsSummary>({
+    queryKey: generateKeys.partyClaimsSummary(caseNumber ?? ""),
+    queryFn: async ({ signal }) => {
+      try {
+        return await apiRequest<{ markdown: string }>(
+          `/api/cases/${caseNumber}/research/party-claims-summary`,
+          { signal },
+        );
+      } catch (e) {
+        if (e instanceof ApiError && e.status === 404) return null;
+        throw e;
+      }
+    },
+    enabled: Boolean(caseNumber),
+    staleTime: 5_000,
+    retry: false,
+  });
+}
+
+/* ── Trigger: party-claims executive summary (button 1) ── */
+export function useGeneratePartyClaimsSummary(caseNumber: string) {
+  const qc = useQueryClient();
+  return useMutation({
+    mutationFn: () =>
+      apiRequest<GenerateTriggerResult>(
+        `/api/cases/${caseNumber}/generate/party-claims-summary`,
+        { method: "POST" },
+      ),
+    onSuccess: () => {
+      // Start polling fresh — the file will appear after the host run completes,
+      // and the status chip should pick up the new queued/running run.
+      qc.invalidateQueries({
+        queryKey: generateKeys.partyClaimsSummary(caseNumber),
+      });
+      qc.invalidateQueries({
+        queryKey: generateKeys.status(caseNumber, "party_claims_summary"),
+      });
+    },
+  });
+}
+
+/* ── Trigger: interim ("party-claims") partial draft (button 2) ── */
+export function useGenerateInterimDraft(caseNumber: string) {
+  const qc = useQueryClient();
+  return useMutation({
+    mutationFn: () =>
+      apiRequest<GenerateTriggerResult>(
+        `/api/cases/${caseNumber}/generate/interim-draft`,
+        { method: "POST" },
+      ),
+    // The result lands in the exports list (טיוטה-טענות_הצדדים_*.docx, polled by
+    // useExports); re-arm the status chip so it shows the new run immediately.
+    onSuccess: () => {
+      qc.invalidateQueries({
+        queryKey: generateKeys.status(caseNumber, "interim_draft"),
+      });
+    },
+  });
+}

web-ui/src/lib/api/legal-arguments.ts

@@ -32,6 +32,12 @@ export type LegalArgument = {
   id: string;
   case_id: string;
   party: LegalArgumentParty;
+  /**
+   * The specific pleading this argument belongs to, for sides that may comprise
+   * several litigants with opposing positions (respondent / permit_applicant
+   * briefs). Empty for single-voice sides (appellant / committee). (#224)
+   */
+  party_name?: string;
   argument_index: number;
   argument_title: string;
   argument_body: string;

web-ui/src/lib/api/operations.ts

@@ -190,6 +190,61 @@ export function useAgentRuns() {
   });
 }
 
+// ── Agent health taxonomy (#222) + recent escalations (#218) ───────────────
+export type AgentHealthState = "zombie" | "stalled" | "working" | "idle";
+
+export type AgentHealthItem = {
+  issue_id: string;
+  identifier: string;
+  status: string;
+  agent_id: string;
+  agent_name: string;
+  health: AgentHealthState;
+  wakeups: number;
+  recovery_wakeups: number;
+};
+
+export type AgentHealthResponse = {
+  ok: boolean;
+  items: AgentHealthItem[]; // worst-first (zombie → stalled → working → idle)
+  counts: Record<AgentHealthState, number>;
+};
+
+/** Per-issue agent health — surfaces recovery-loop zombies automatically. */
+export function useAgentHealth() {
+  return useQuery({
+    queryKey: ["operations", "agents", "health"],
+    queryFn: ({ signal }) =>
+      apiRequest<AgentHealthResponse>("/api/operations/agents/health", { signal }),
+    refetchInterval: 5000, // live alongside the running-agents view
+    staleTime: 3000,
+  });
+}
+
+export type EscalationItem = {
+  issue_id: string;
+  identifier: string;
+  severity: "critical" | "high" | "medium";
+  reason: string;
+  created_at: string | null;
+};
+
+export type RecentEscalationsResponse = {
+  ok: boolean;
+  items: EscalationItem[]; // newest-first
+};
+
+/** Recent chair escalations (watchdog + manual) — the handled-loop history. */
+export function useRecentEscalations() {
+  return useQuery({
+    queryKey: ["operations", "agents", "escalations"],
+    queryFn: ({ signal }) =>
+      apiRequest<RecentEscalationsResponse>("/api/operations/agents/escalations", { signal }),
+    refetchInterval: 15000, // escalations change slowly (watchdog runs every 15m)
+    staleTime: 10000,
+  });
+}
+
 /** Full output log of one run — fetched on demand (drawer open). */
 export function useRunLog(runId: string | null) {
   return useQuery({

web-ui/src/lib/api/protocol-analysis.ts

@@ -0,0 +1,134 @@
+/**
+ * Protocol comparative analysis — the "מה קרה בדיון" panel (#226).
+ *
+ * `analyze_protocol` compares the ערר hearing protocol against the written
+ * pleadings and records, per argument, whether it was strengthened, newly
+ * raised, or dropped at the hearing — plus the sharpened legal question and a
+ * verbatim evidence quote. This module exposes the read + trigger endpoints.
+ */
+
+import { useMutation, useQuery, useQueryClient } from "@tanstack/react-query";
+import { apiRequest } from "./client";
+import type { LegalArgumentParty } from "./legal-arguments";
+
+export type ProtocolChangeType = "strengthened" | "newly_raised" | "dropped";
+
+/** Empty string = an unassigned party_role on a verdict row. */
+export type ProtocolPartyRole = LegalArgumentParty | "";
+
+export type ProtocolAnalysisRow = {
+  id: string;
+  case_id: string;
+  document_id: string;
+  party_role: ProtocolPartyRole;
+  change_type: ProtocolChangeType;
+  /** The pleaded argument this verdict matched (strengthened/dropped); null for newly_raised. */
+  argument_id: string | null;
+  argument_title: string;
+  summary: string;
+  sharpened_question: string;
+  evidence_quote: string;
+  page_number: number | null;
+  created_at?: string;
+};
+
+export type ProtocolHeader = {
+  hearing_date: string | null;
+  protocol_title: string;
+  protocol_document_id: string;
+  panel_members: string[];
+  appellants_present: string[];
+  respondents_present: string[];
+};
+
+export type ProtocolAnalysisResponse = {
+  case_number: string;
+  total: number;
+  header: ProtocolHeader;
+  by_change: Partial<Record<ProtocolChangeType, number>>;
+  by_party: Partial<Record<ProtocolPartyRole, ProtocolAnalysisRow[]>>;
+  analysis: ProtocolAnalysisRow[];
+};
+
+export const protocolAnalysisKeys = {
+  all: ["protocol-analysis"] as const,
+  byCase: (caseNumber: string) =>
+    [...protocolAnalysisKeys.all, caseNumber] as const,
+};
+
+export function useProtocolAnalysis(caseNumber: string | undefined) {
+  return useQuery({
+    queryKey: protocolAnalysisKeys.byCase(caseNumber ?? ""),
+    queryFn: ({ signal }) =>
+      apiRequest<ProtocolAnalysisResponse>(
+        `/api/cases/${caseNumber}/protocol-analysis`,
+        { signal },
+      ),
+    enabled: Boolean(caseNumber),
+    staleTime: 10_000,
+  });
+}
+
+/**
+ * The analysis runs on the legal-analyst agent (host-side, where the `claude`
+ * CLI lives) — NOT inline in the FastAPI container. The endpoint delegates via
+ * a Paperclip wakeup (`queued`), or reports `skipped` when no analyst route is
+ * available (the chair can then run the MCP tool manually).
+ */
+export type AnalyzeProtocolResult =
+  | {
+      status: "queued";
+      sub_issue_id: string;
+      analyst_id: string;
+      main_issue_id: string;
+    }
+  | {
+      status: "skipped";
+      reason: "no_api_key" | "no_analyst" | "no_issue" | string;
+      company_id?: string;
+    };
+
+export function useAnalyzeProtocol(caseNumber: string | undefined) {
+  const qc = useQueryClient();
+  return useMutation({
+    mutationFn: (documentId?: string) =>
+      apiRequest<AnalyzeProtocolResult>(
+        `/api/cases/${caseNumber}/analyze-protocol${
+          documentId ? `?document_id=${encodeURIComponent(documentId)}` : ""
+        }`,
+        { method: "POST" },
+      ),
+    onSuccess: () => {
+      if (caseNumber) {
+        qc.invalidateQueries({
+          queryKey: protocolAnalysisKeys.byCase(caseNumber),
+        });
+      }
+    },
+  });
+}
+
+export const CHANGE_LABELS_HE: Record<ProtocolChangeType, string> = {
+  strengthened: "התחזק בדיון",
+  newly_raised: "נטען לראשונה",
+  dropped: "נזנח בדיון",
+};
+
+export const PROTOCOL_PARTY_LABELS_HE: Record<ProtocolPartyRole, string> = {
+  appellant: "עוררים",
+  respondent: "משיבים",
+  committee: "ועדה מקומית",
+  permit_applicant: "מבקשי היתר",
+  unknown: "צד לא מזוהה",
+  "": "צד לא מסווג",
+};
+
+/** Display order for the per-party sections. */
+export const PROTOCOL_PARTY_ORDER: ProtocolPartyRole[] = [
+  "appellant",
+  "committee",
+  "respondent",
+  "permit_applicant",
+  "unknown",
+  "",
+];

web-ui/src/lib/doc-types.ts

@@ -64,6 +64,79 @@ export function doctypeTone(value: string): string {
   }
 }
 
+// ── Overview groups (display-only convenience buckets) ─────────────
+//
+// The case-overview document list is split into four sections purely for
+// the chair's convenience — grouping NEVER changes a document's stored
+// doc_type. Classification is derived on the client from doc_type plus two
+// metadata flags (is_post_hearing, protocol_scope).
+
+export type DocGroup = "primary" | "ancillary" | "hearing_protocol" | "post_hearing";
+
+export const DOC_GROUP_LABELS: Record<DocGroup, string> = {
+  primary:          "מסמכים עיקריים",
+  ancillary:        "מסמכים נלווים",
+  hearing_protocol: "פרוטוקול דיון ועדת הערר",
+  post_hearing:     "מסמכים לאחר הדיון",
+};
+
+export const DOC_GROUP_HINTS: Record<DocGroup, string> = {
+  primary:          "כתבי ערר, תשובות ושומות שהוגשו לפני הדיון",
+  ancillary:        "פרוטוקולי/החלטות ועדה מקומית-מחוזית, תכניות, וכל השאר",
+  hearing_protocol: "הדיון בפני הוועדה",
+  post_hearing:     "השלמות ותגובות שהתקבלו לאחר הדיון",
+};
+
+/** Fixed render order of the groups in the overview panel. */
+export const DOC_GROUP_ORDER: DocGroup[] = [
+  "primary",
+  "ancillary",
+  "hearing_protocol",
+  "post_hearing",
+];
+
+/** doc_types that count as "primary" pre-hearing filings (group 1). */
+const PRIMARY_DOC_TYPES = new Set<DocType>(["appeal", "response", "appraisal"]);
+
+/**
+ * Classify a document into one of the four overview groups.
+ *
+ * Priority order matters:
+ *   1. is_post_hearing wins over everything — a post-hearing כתב תשובה
+ *      belongs in group 4, not group 1.
+ *   2. A protocol goes to group 3 (ועדת הערר) by default, or group 2 when
+ *      explicitly tagged as a lower-committee protocol (protocol_scope==="lower").
+ *   3. appeal/response/appraisal → group 1; everything else → group 2.
+ */
+export function documentGroup(
+  docType: string,
+  metadata?: Record<string, unknown> | null,
+): DocGroup {
+  if (metadata?.is_post_hearing === true) return "post_hearing";
+  if (docType === "protocol") {
+    return metadata?.protocol_scope === "lower" ? "ancillary" : "hearing_protocol";
+  }
+  if (PRIMARY_DOC_TYPES.has(docType as DocType)) return "primary";
+  return "ancillary";
+}
+
+// ── Protocol scope (only relevant when doc_type === "protocol") ────
+// Distinguishes the appeal-committee hearing protocol (group 3) from a
+// lower (local/district) committee protocol (group 2). Stored in
+// documents.metadata.protocol_scope; absent === "appeal" (the default).
+
+export type ProtocolScope = "appeal" | "lower";
+
+export const PROTOCOL_SCOPE_LABELS: Record<ProtocolScope, string> = {
+  appeal: "ועדת הערר",
+  lower:  "ועדה מקומית/מחוזית",
+};
+
+export const PROTOCOL_SCOPE_OPTIONS: { value: ProtocolScope; label: string }[] = [
+  { value: "appeal", label: PROTOCOL_SCOPE_LABELS.appeal },
+  { value: "lower",  label: PROTOCOL_SCOPE_LABELS.lower },
+];
+
 // ── Appraiser sides (only relevant when doc_type === "appraisal") ──
 
 export type AppraiserSide = "committee" | "appellant" | "deciding";

web-ui/src/lib/format-date.ts

@@ -0,0 +1,296 @@
+/**
+ * Central date/time formatting — pinned to Israel time (Asia/Jerusalem).
+ *
+ * WHY THIS EXISTS (INV-UI9 / G2):
+ * Storage stays UTC end-to-end (Postgres TIMESTAMPTZ, API returns ISO-UTC).
+ * Only the *display* layer is localized, and it is localized **deterministically**:
+ * every human-facing timestamp renders in `Asia/Jerusalem`, regardless of the
+ * host/container/browser timezone. The legal-ai container runs in UTC, so an
+ * ad-hoc `new Date(iso).toLocaleDateString("he-IL")` shows UTC under SSR (and
+ * UTC again on any non-Israel runtime) — that drift is exactly what this module
+ * removes. Pinning `timeZone: "Asia/Jerusalem"` makes SSR and the browser agree.
+ *
+ * This is the ONE date formatter for the UI. Do not format timestamps ad-hoc
+ * with `new Date(...).toLocaleDateString(...)` / `.toLocaleString(...)` /
+ * `.getHours()` rendering elsewhere — route through these helpers so the tz is
+ * always Israel and nothing drifts.
+ *
+ * Number formatting (e.g. `n.toLocaleString("he-IL")`) is NOT a timestamp and is
+ * intentionally out of scope here.
+ */
+
+const TZ = "Asia/Jerusalem";
+const LOCALE = "he-IL";
+
+/** Coerce an ISO string / epoch-ms / Date into a valid Date, or null. */
+function toDate(value: string | number | Date | null | undefined): Date | null {
+  if (value == null) return null;
+  const d = value instanceof Date ? value : new Date(value);
+  return Number.isNaN(d.getTime()) ? null : d;
+}
+
+/**
+ * Date only — `dd/MM/yyyy` in Israel time (e.g. "30/06/2026").
+ * Matches the prior `toLocaleDateString("he-IL", {day,month:"2-digit",year:"numeric"})`.
+ */
+export function formatDate(
+  value: string | number | Date | null | undefined,
+  fallback = "—",
+): string {
+  const d = toDate(value);
+  if (!d) return fallback;
+  return new Intl.DateTimeFormat(LOCALE, {
+    timeZone: TZ,
+    day: "2-digit",
+    month: "2-digit",
+    year: "numeric",
+  }).format(d);
+}
+
+/**
+ * Date only — long month form (e.g. "30 ביוני 2026") in Israel time.
+ * Matches the prior `toLocaleDateString("he-IL", {day:"numeric",month:"long",year:"numeric"})`.
+ */
+export function formatDateLong(
+  value: string | number | Date | null | undefined,
+  fallback = "",
+): string {
+  const d = toDate(value);
+  if (!d) return fallback;
+  return new Intl.DateTimeFormat(LOCALE, {
+    timeZone: TZ,
+    day: "numeric",
+    month: "long",
+    year: "numeric",
+  }).format(d);
+}
+
+/**
+ * Date only — short form, locale default (e.g. "30.6.2026") in Israel time.
+ * Matches the prior bare `toLocaleDateString("he-IL")`.
+ */
+export function formatDateShort(
+  value: string | number | Date | null | undefined,
+  fallback = "—",
+): string {
+  const d = toDate(value);
+  if (!d) return fallback;
+  return new Intl.DateTimeFormat(LOCALE, { timeZone: TZ }).format(d);
+}
+
+/**
+ * Date + time — `d בMMM yyyy, HH:mm` style in Israel time.
+ * Matches the prior `{day:"numeric",month:"short",year:"numeric",hour,minute}`.
+ */
+export function formatDateTime(
+  value: string | number | Date | null | undefined,
+  fallback = "—",
+): string {
+  const d = toDate(value);
+  if (!d) return fallback;
+  return new Intl.DateTimeFormat(LOCALE, {
+    timeZone: TZ,
+    day: "numeric",
+    month: "short",
+    year: "numeric",
+    hour: "2-digit",
+    minute: "2-digit",
+  }).format(d);
+}
+
+/**
+ * Full date + time — locale default with time, Israel time.
+ * Matches the prior bare `toLocaleString("he-IL")`.
+ */
+export function formatDateTimeFull(
+  value: string | number | Date | null | undefined,
+  fallback = "—",
+): string {
+  const d = toDate(value);
+  if (!d) return fallback;
+  return new Intl.DateTimeFormat(LOCALE, {
+    timeZone: TZ,
+    dateStyle: "short",
+    timeStyle: "short",
+  }).format(d);
+}
+
+/**
+ * Time only — `HH:mm` in Israel time (e.g. "14:05").
+ * Matches the prior `toLocaleTimeString("he-IL", {hour,minute})`.
+ */
+export function formatTime(
+  value: string | number | Date | null | undefined,
+  fallback = "—",
+): string {
+  const d = toDate(value);
+  if (!d) return fallback;
+  return new Intl.DateTimeFormat(LOCALE, {
+    timeZone: TZ,
+    hour: "2-digit",
+    minute: "2-digit",
+  }).format(d);
+}
+
+/**
+ * ISO calendar date `YYYY-MM-DD` for the Israel-time day of the instant.
+ * Deterministic replacement for `new Date(iso).toISOString().slice(0,10)`,
+ * which yields the UTC day (off-by-one near midnight Israel time).
+ */
+export function formatIsoDate(
+  value: string | number | Date | null | undefined,
+  fallback = "—",
+): string {
+  const d = toDate(value);
+  if (!d) return fallback;
+  // en-CA's short date is ISO-shaped (YYYY-MM-DD); pin the tz so it's the Israel day.
+  return new Intl.DateTimeFormat("en-CA", {
+    timeZone: TZ,
+    year: "numeric",
+    month: "2-digit",
+    day: "2-digit",
+  }).format(d);
+}
+
+/**
+ * The calendar year (number) of the Israel-time day for an instant.
+ * Deterministic replacement for `new Date(iso).getFullYear()` (which uses the
+ * runtime tz). Returns null when the value is absent/unparseable.
+ */
+export function getIsraelYear(
+  value: string | number | Date | null | undefined,
+): number | null {
+  const d = toDate(value);
+  if (!d) return null;
+  const y = new Intl.DateTimeFormat("en-CA", { timeZone: TZ, year: "numeric" }).format(d);
+  const n = Number(y);
+  return Number.isNaN(n) ? null : n;
+}
+
+/**
+ * The Israel-time calendar day of an instant as a sortable integer YYYYMMDD,
+ * or null when absent/unparseable. Use this to compare two dates at *day*
+ * precision in Israel time (e.g. hearing-date vs. today) without runtime-tz drift.
+ */
+export function israelDayKey(
+  value: string | number | Date | null | undefined,
+): number | null {
+  const iso = formatIsoDate(value, ""); // "YYYY-MM-DD" or ""
+  if (!iso) return null;
+  const n = Number(iso.replace(/-/g, ""));
+  return Number.isNaN(n) ? null : n;
+}
+
+/** Today's Israel-time calendar day as a sortable integer YYYYMMDD. */
+export function todayIsraelDayKey(): number {
+  return israelDayKey(new Date()) as number;
+}
+
+/**
+ * The UTC-ms of midnight (00:00) of the instant's *Israel* calendar day.
+ * Day-precision pivot for ms-based comparisons/diffs (e.g. hearing date vs.
+ * today) that must not drift with the runtime timezone. Returns null when the
+ * value is absent/unparseable. Built from the en-CA YYYY-MM-DD of the Israel day
+ * combined with the Israel UTC offset at that instant, so two values from the
+ * same Israel day always reduce to the same ms.
+ */
+export function israelMidnightMs(
+  value: string | number | Date | null | undefined,
+): number | null {
+  const d = toDate(value);
+  if (!d) return null;
+  const iso = formatIsoDate(d, ""); // YYYY-MM-DD of the Israel day
+  if (!iso) return null;
+  const [y, m, day] = iso.split("-").map(Number);
+  // Israel wall-clock midnight as a UTC instant: take naive UTC midnight of that
+  // calendar day and subtract Israel's offset (Israel is ahead of UTC).
+  const naiveUtcMidnight = Date.UTC(y, m - 1, day);
+  const utc = new Date(d.toLocaleString("en-US", { timeZone: "UTC" }));
+  const isr = new Date(d.toLocaleString("en-US", { timeZone: TZ }));
+  const offsetMs = isr.getTime() - utc.getTime();
+  return naiveUtcMidnight - offsetMs;
+}
+
+/** UTC-ms of midnight of *today's* Israel calendar day. */
+export function todayIsraelMidnightMs(): number {
+  return israelMidnightMs(new Date()) as number;
+}
+
+/**
+ * Numeric date/time components of an instant in Israel time. For the rare call
+ * site that builds a bespoke string from parts (e.g. a custom weekday + dd/MM
+ * HH:mm deadline label) and must do so in Israel time rather than the runtime
+ * tz. `weekday` is 0=Sunday … 6=Saturday (matching `Date.getDay()`).
+ * Returns null when the value is absent/unparseable.
+ */
+export function israelParts(
+  value: string | number | Date | null | undefined,
+): { year: number; month: number; day: number; hour: number; minute: number; weekday: number } | null {
+  const d = toDate(value);
+  if (!d) return null;
+  const parts = new Intl.DateTimeFormat("en-US", {
+    timeZone: TZ,
+    year: "numeric",
+    month: "2-digit",
+    day: "2-digit",
+    hour: "2-digit",
+    minute: "2-digit",
+    weekday: "short",
+    hour12: false,
+  }).formatToParts(d);
+  const get = (t: string) => parts.find((p) => p.type === t)?.value ?? "";
+  const WD: Record<string, number> = { Sun: 0, Mon: 1, Tue: 2, Wed: 3, Thu: 4, Fri: 5, Sat: 6 };
+  let hour = Number(get("hour"));
+  if (hour === 24) hour = 0; // some engines emit "24" for midnight under hour12:false
+  return {
+    year: Number(get("year")),
+    month: Number(get("month")),
+    day: Number(get("day")),
+    hour,
+    minute: Number(get("minute")),
+    weekday: WD[get("weekday")] ?? 0,
+  };
+}
+
+/**
+ * Relative "time ago" in Hebrew. Timezone-agnostic (it's a diff between two
+ * absolute instants), but centralized here so all relative rendering lives in
+ * one place. `units` selects the existing wording per call site so visible text
+ * is unchanged:
+ *   - "long":  עכשיו / לפני N דקות / לפני N שעות / לפני N ימים
+ *   - "short": עכשיו / לפני N דק׳ / לפני N שע׳ / לפני N ימים
+ *   - "tight": לפני Nד' / לפני Nש' / לפני N ימים  (no "now" bucket)
+ */
+export function formatRelative(
+  value: string | number | Date | null | undefined,
+  opts: { units?: "long" | "short" | "tight"; fallback?: string } = {},
+): string {
+  const { units = "short", fallback = "" } = opts;
+  const d = toDate(value);
+  if (!d) return fallback;
+
+  const diffMs = Date.now() - d.getTime();
+  const mins = Math.floor(diffMs / 60_000);
+  const hours = Math.floor(mins / 60);
+  const days = Math.floor(hours / 24);
+
+  if (units === "tight") {
+    if (mins < 60) return `לפני ${mins}ד'`;
+    if (hours < 24) return `לפני ${hours}ש'`;
+    return `לפני ${days} ימים`;
+  }
+
+  const now = "עכשיו";
+  if (units === "long") {
+    if (mins < 1) return now;
+    if (mins < 60) return `לפני ${mins} דקות`;
+    if (hours < 24) return `לפני ${hours} שעות`;
+    return `לפני ${days} ימים`;
+  }
+
+  // "short"
+  if (mins < 1) return now;
+  if (mins < 60) return `לפני ${mins} דק׳`;
+  if (hours < 24) return `לפני ${hours} שע׳`;
+  return `לפני ${days} ימים`;
+}

web/agent_health.py

@@ -0,0 +1,110 @@
+"""Agent health taxonomy (TaskMaster #222) — one crisp state per stuck issue.
+
+Classifies each open, agent-assigned issue into a single health state so the
+stranded-child / recovery-loop cases we diagnose by hand today
+([[reference_recovery_loop_stranded_child]], [[reference_paperclip_recovery_loops]])
+surface automatically in the dashboard — the ``gt feed --problems`` idea from
+Gastown, grounded in *our* failure modes.
+
+Two layers, split for testability (same shape as web.agent_telemetry):
+- :func:`classify_issue_health` is a **pure function** over three primitives
+  (live-run flag + wakeup counts). No DB, no platform symbols — trivially
+  unit-tested and platform-agnostic (INV-G12).
+- The Paperclip-specific fetch that derives those primitives from
+  ``heartbeat_runs`` / ``agent_wakeup_requests`` lives in the shell
+  (``web.paperclip_client.get_agent_health``); this module only decides *how to
+  label* what the fetch measured, so the taxonomy survives a platform swap.
+
+The wakeup/recovery signals it reads are the very ones #219 telemetry now emits
+(``agent.wakeup`` with recovery reasons) — this is the consumer that turns that
+stream into an at-a-glance verdict.
+"""
+from __future__ import annotations
+
+# Severity order (worst first) — the dashboard sorts problems by this.
+HEALTH_STATES: tuple[str, ...] = ("zombie", "stalled", "working", "idle")
+
+# Substrings of Paperclip wakeup ``reason`` values that mark a recovery-loop
+# wake (as opposed to a genuine new task). Sourced from the recovery-loop
+# forensics in memory: an issue that keeps getting these while agent-owned is a
+# zombie — assigned + "active" but nothing is really executing.
+RECOVERY_REASON_MARKERS: tuple[str, ...] = (
+    "source_scoped_recovery",
+    "stranded_assigned",
+    "issue_continuation",
+    "issue_reopened",
+    "missing_disposition",
+    "successful_run_missing_state",
+    "waiting_on_review",
+)
+
+# ≥ this many wakeups inside the fetch window, with no recovery marker and no
+# live run, reads as "poked repeatedly but not progressing" = stalled.
+STALL_WAKEUP_THRESHOLD = 3
+
+
+def is_recovery_reason(reason: str | None) -> bool:
+    """True if a wakeup ``reason`` is a recovery-loop marker (not a new task)."""
+    if not reason:
+        return False
+    return any(marker in reason for marker in RECOVERY_REASON_MARKERS)
+
+
+def classify_issue_health(
+    *,
+    has_live_run: bool,
+    recovery_wakeups: int,
+    total_wakeups: int,
+) -> str:
+    """Label one open, agent-assigned issue. Pure — no I/O.
+
+    - ``working`` — a live run is executing right now.
+    - ``zombie`` — no live run but recovery-loop wakeups are firing: the issue
+      looks agent-owned + active yet nothing is really running (the stranded
+      child we chase manually).
+    - ``stalled`` — no live run, no recovery markers, but poked ``>=``
+      :data:`STALL_WAKEUP_THRESHOLD` times in-window (repeated wakes, no
+      progress).
+    - ``idle`` — open + assigned but quiet: benign waiting.
+
+    Called only for issues that are already open and agent-assigned; unassigned
+    or closed issues are not health-tracked (the fetch filters them out).
+    """
+    if has_live_run:
+        return "working"
+    if recovery_wakeups >= 1:
+        return "zombie"
+    if total_wakeups >= STALL_WAKEUP_THRESHOLD:
+        return "stalled"
+    return "idle"
+
+
+# A zombie needs at least this many recovery-marker wakeups before the watchdog
+# auto-escalates it — filters a single transient recovery wake from a genuine,
+# budget-burning loop. Escalation reassigns the issue to the chair (via
+# escalate_issue), so an escalated issue is agent-unowned next sweep and never
+# re-escalates (idempotent by construction).
+ZOMBIE_ESCALATE_MIN_RECOVERY = 2
+
+
+def zombie_escalation(item: dict) -> tuple[str, str] | None:
+    """Decide whether a health item warrants auto-escalation to the chair.
+
+    Pure (no I/O) so the watchdog's policy is unit-tested. Returns
+    ``(severity, reason)`` for a persistent zombie, else ``None``. Severity
+    scales with loop intensity — a long-running loop (``>=5`` recovery wakes)
+    is ``high``, a fresh one ``medium``. Only ``zombie`` items escalate: a
+    ``stalled``/``working``/``idle`` issue is not a recovery loop.
+    """
+    if item.get("health") != "zombie":
+        return None
+    recovery = int(item.get("recovery_wakeups", 0))
+    if recovery < ZOMBIE_ESCALATE_MIN_RECOVERY:
+        return None
+    severity = "high" if recovery >= 5 else "medium"
+    reason = (
+        f"auto-escalation (watchdog): זוהתה לולאת-recovery — {recovery} יקיצות-שחזור "
+        f"על {item.get('identifier')} ({item.get('agent_name')}); ה-issue נמסר לחיים "
+        f"במקום להמשיך לשרוף budget."
+    )
+    return severity, reason

web/agent_platform_port.py

@@ -33,37 +33,118 @@ from web.paperclip_api import (
     pc_request,
     require_paperclip_db_url,
 )
+from web.agent_telemetry import instrument
 from web.paperclip_client import (
     COMPANIES as PAPERCLIP_COMPANIES,
     accept_interaction as pc_accept_interaction,
     archive_project as pc_archive_project,
-    cancel_run as pc_cancel_run,
     create_project as pc_create_project,
     create_workflow_issue as pc_create_workflow_issue,
+    get_agent_health as pc_get_agent_health,
+    get_recent_escalations as pc_get_recent_escalations,
+    get_predecessor_context as pc_get_predecessor_context,
+    get_predecessor_for_case as pc_get_predecessor_for_case,
     get_agents_for_case as pc_get_agents_for_case,
     get_agents_for_company as pc_get_agents,
     get_case_issues as pc_get_case_issues,
+    pick_default_comment_target as pc_pick_default_comment_target,
     get_issue_comments as pc_get_issue_comments,
     get_issue_interactions as pc_get_issue_interactions,
     get_project_url,
     get_run_events as pc_get_run_events,
     get_run_log as pc_get_run_log,
     list_live_runs as pc_list_live_runs,
-    post_comment as pc_post_comment,
     reject_interaction as pc_reject_interaction,
-    reset_agent_session as pc_reset_agent_session,
-    reset_case_agents as pc_reset_case_agents,
     respond_to_interaction as pc_respond_to_interaction,
     restore_project as pc_restore_project,
     update_project_name as pc_update_project_name,
-    wake_analyst_for_appraiser_facts as pc_wake_analyst_for_appraiser_facts,
-    wake_analyst_for_argument_aggregation as pc_wake_analyst_for_argument_aggregation,
-    wake_ceo_agent as pc_wake_ceo,
-    wake_ceo_for_feedback_fold as pc_wake_ceo_for_feedback_fold,
-    wake_curator_for_final as pc_wake_curator_for_final,
-    wake_for_precedent_extraction as pc_wake_for_precedent_extraction,
+    get_generation_run_status as pc_get_generation_run_status,
+    # ── raw imports: wrapped with telemetry below (state-affecting ops, #219) ──
+    cancel_interaction as _cancel_interaction,
+    cancel_run as _cancel_run,
+    escalate_issue as _escalate_issue,
+    mark_comment_routed as _mark_comment_routed,
+    post_comment as _post_comment,
+    reap_stale_interactions as _reap_stale_interactions,
+    reset_agent_session as _reset_agent_session,
+    reset_case_agents as _reset_case_agents,
+    set_issue_status as _set_issue_status,
+    wake_analyst_for_appraiser_facts as _wake_analyst_for_appraiser_facts,
+    wake_analyst_for_argument_aggregation as _wake_analyst_for_argument_aggregation,
+    wake_analyst_for_protocol_analysis as _wake_analyst_for_protocol_analysis,
+    wake_ceo_agent as _wake_ceo,
+    wake_ceo_for_action as _wake_ceo_for_action,
+    open_ceo_run as _open_ceo_run,
+    wake_ceo_for_feedback_fold as _wake_ceo_for_feedback_fold,
+    wake_curator_for_final as _wake_curator_for_final,
+    wake_for_precedent_extraction as _wake_for_precedent_extraction,
 )
 
+# ── telemetry-wrapped platform ops (#219 / docs/spec/X15) ───────────────────
+# Every state-affecting platform op the Port exposes emits one structured event
+# through web.agent_telemetry, so recovery-loop behaviour (repeated wakeups,
+# stranded dispositions) is observable in real time rather than reconstructed
+# from the Paperclip DB. The wrappers preserve the public ``pc_*`` names and
+# signatures — call sites in app.py are unchanged. Read-only observability ops
+# (list_live_runs / get_run_log / get_issue_comments …) are intentionally NOT
+# instrumented: they neither wake agents nor change disposition state.
+pc_wake_ceo = instrument("agent.wakeup", role="ceo")(_wake_ceo)
+pc_wake_ceo_for_action = instrument(
+    "agent.wakeup", role="ceo", keys=("case_number", "company_id", "action"),
+)(_wake_ceo_for_action)
+pc_open_ceo_run = instrument(
+    "agent.wakeup", role="ceo", keys=("case_number", "company_id"),
+)(_open_ceo_run)
+pc_wake_ceo_for_feedback_fold = instrument(
+    "agent.wakeup", role="ceo", keys=("feedback_id", "category", "block_id"),
+)(_wake_ceo_for_feedback_fold)
+pc_wake_curator_for_final = instrument(
+    "agent.wakeup", role="curator", keys=("case_number", "company_id", "task"),
+)(_wake_curator_for_final)
+pc_wake_for_precedent_extraction = instrument(
+    "agent.wakeup", role="ceo", keys=("case_law_id", "citation", "practice_area"),
+)(_wake_for_precedent_extraction)
+pc_wake_analyst_for_appraiser_facts = instrument(
+    "agent.wakeup", role="analyst",
+)(_wake_analyst_for_appraiser_facts)
+pc_wake_analyst_for_argument_aggregation = instrument(
+    "agent.wakeup", role="analyst",
+)(_wake_analyst_for_argument_aggregation)
+pc_wake_analyst_for_protocol_analysis = instrument(
+    "agent.wakeup", role="analyst", keys=("case_number", "company_id", "document_id"),
+)(_wake_analyst_for_protocol_analysis)
+pc_post_comment = instrument(
+    "agent.comment", keys=("issue_id", "company_id"),
+)(_post_comment)
+pc_mark_comment_routed = instrument(
+    "agent.comment", keys=("issue_id", "comment_id"),
+)(_mark_comment_routed)
+pc_cancel_interaction = instrument(
+    "interaction.cancelled", keys=("issue_id", "interaction_id"),
+)(_cancel_interaction)
+# First-class escalation-to-chair (#218): the loop-safe alternative to leaving an
+# issue agent-owned+blocked. Emits a severity-tagged event so escalations are
+# countable per case in the same stream as the wakeups they replace.
+pc_escalate_issue = instrument(
+    "agent.escalated", keys=("issue_id", "severity", "company_id", "reason"),
+)(_escalate_issue)
+# Manual chair close of a board issue (done/cancelled) — the loop-safe primitive
+# behind the agents-board "סמן כהושלם / בטל" actions. Countable per case in the
+# same telemetry stream as the escalations/wakeups it sits beside.
+pc_set_issue_status = instrument(
+    "issue.status_set", keys=("issue_id", "status", "company_id"),
+)(_set_issue_status)
+pc_reap_stale_interactions = instrument(
+    "interaction.reaped", result_keys=("cancelled",),
+)(_reap_stale_interactions)
+pc_cancel_run = instrument("run.cancelled", keys=("run_id",))(_cancel_run)
+pc_reset_agent_session = instrument(
+    "agent.session_reset", keys=("agent_id",),
+)(_reset_agent_session)
+pc_reset_case_agents = instrument(
+    "agent.session_reset", keys=("case_number",),
+)(_reset_case_agents)
+
 # ── domain-named lifecycle aliases (preferred for new call sites) ───────────
 archive_case_project = pc_archive_project
 restore_case_project = pc_restore_project
@@ -93,23 +174,37 @@ __all__ = [
     # issues / workflow
     "pc_create_workflow_issue",
     "pc_get_case_issues",
+    "pc_pick_default_comment_target",
     # agents / wakeups
     "pc_get_agents_for_case",
     "pc_get_agents",
     "pc_wake_ceo",
+    "pc_wake_ceo_for_action",
+    "pc_open_ceo_run",
     "pc_wake_ceo_for_feedback_fold",
     "pc_wake_curator_for_final",
     "pc_wake_for_precedent_extraction",
     "pc_wake_analyst_for_appraiser_facts",
     "pc_wake_analyst_for_argument_aggregation",
+    "pc_wake_analyst_for_protocol_analysis",
+    "pc_get_generation_run_status",
     # comments / interactions
     "pc_post_comment",
+    "pc_mark_comment_routed",
     "pc_get_issue_comments",
     "pc_get_issue_interactions",
     "pc_accept_interaction",
     "pc_reject_interaction",
     "pc_respond_to_interaction",
+    "pc_cancel_interaction",
+    "pc_reap_stale_interactions",
+    "pc_escalate_issue",
+    "pc_set_issue_status",
     # agent-run observability + control (live view + smart management)
+    "pc_get_agent_health",
+    "pc_get_recent_escalations",
+    "pc_get_predecessor_context",
+    "pc_get_predecessor_for_case",
     "pc_list_live_runs",
     "pc_get_run_log",
     "pc_get_run_events",

web/agent_telemetry.py

@@ -0,0 +1,157 @@
+"""Agent-platform telemetry (TaskMaster #219) — structured, platform-agnostic events.
+
+Emits one structured event per agent-platform operation (wakeup, disposition
+comment, interaction reap/cancel, run cancel, session reset) so recovery-loop
+behaviour is **observable in real time** instead of reconstructed forensically
+from the Paperclip DB (``agent_wakeup_requests`` / ``heartbeat_runs``). See
+[[reference_paperclip_recovery_loops]] / [[reference_recovery_loop_stranded_child]]
+for the failure modes this makes visible.
+
+**Platform-agnostic by design (INV-G12 / docs/spec/X15).** Events are named in
+domain terms and carry no Paperclip-specific symbols — this module imports
+*nothing* from the Paperclip shell. The Port (``web/agent_platform_port.py``)
+calls :func:`instrument` / :func:`emit` around its operations, so the telemetry
+semantics survive a platform swap.
+
+**OTLP-ready, "logs first" (#219).** Each event is a flat dict with a stable
+schema, written today as one JSON line on the ``agent.telemetry`` logger
+(greppable in Coolify logs). Swapping :data:`_SINK` for an OTLP exporter — or
+attaching an OTLP handler to that logger — is the later step, with **no
+call-site changes**. Timestamps are stored in UTC (display-in-Israel is a
+read-side concern; see [[feedback_israel_time]]).
+
+Single path (INV-G2): the only telemetry emitter for the agent platform. Do not
+add a parallel structured-log format for these events elsewhere — route through
+:func:`emit`.
+"""
+from __future__ import annotations
+
+import functools
+import inspect
+import json
+import logging
+import time
+from collections.abc import Awaitable, Callable, Sequence
+from datetime import datetime, timezone
+from typing import Any, TypeVar
+
+logger = logging.getLogger("agent.telemetry")
+
+# Schema version — bump when the event shape changes so downstream consumers
+# (dashboards, the #222 health taxonomy) can migrate deliberately.
+SCHEMA_VERSION = 1
+
+# Standard argument keys lifted onto every event when present on the wrapped call.
+_STANDARD_KEYS: tuple[str, ...] = ("case_number", "issue_id", "company_id", "reason")
+
+T = TypeVar("T")
+
+
+def _now_iso() -> str:
+    """UTC ISO-8601 with a trailing ``Z`` — store-UTC (feedback_israel_time)."""
+    return datetime.now(timezone.utc).isoformat(timespec="milliseconds").replace("+00:00", "Z")
+
+
+def _sink(payload: dict[str, Any]) -> None:
+    """Where an event goes. Today: a JSON line on the ``agent.telemetry`` logger.
+
+    Isolated so a future OTLP exporter is a one-line swap with no call-site
+    churn (#219 "backend later").
+    """
+    logger.info(json.dumps(payload, ensure_ascii=False, default=str))
+
+
+_SINK: Callable[[dict[str, Any]], None] = _sink
+
+
+def emit(event: str, *, outcome: str = "ok", **fields: Any) -> None:
+    """Emit one structured agent-platform event. Never raises.
+
+    Telemetry must never break the operation it observes — any failure in the
+    sink is swallowed to a warning. ``event`` is a dotted domain name
+    (``agent.wakeup``, ``interaction.reaped``, ``run.cancelled`` …). Extra
+    ``fields`` are merged flat into the event; ``None`` values are dropped to
+    keep events terse.
+    """
+    payload: dict[str, Any] = {
+        "ts": _now_iso(),
+        "v": SCHEMA_VERSION,
+        "event": event,
+        "outcome": outcome,
+    }
+    for key, value in fields.items():
+        if value is not None:
+            payload[key] = value
+    try:
+        _SINK(payload)
+    except Exception:  # pragma: no cover — telemetry is best-effort, never fatal
+        logger.warning("agent.telemetry emit failed for event=%s", event, exc_info=True)
+
+
+def _outcome_of(result: Any) -> str:
+    """Derive an outcome from a platform call's return value.
+
+    Our platform helpers return either a plain dict (Paperclip API JSON) or a
+    ``{"ok": bool, ...}`` envelope. A falsy ``ok`` is a *no-op* (e.g. an
+    interaction that was already resolved), distinct from a raised error.
+    """
+    if isinstance(result, dict) and result.get("ok") is False:
+        return "noop"
+    return "ok"
+
+
+def _duration_ms(t0: float) -> int:
+    return int((time.monotonic() - t0) * 1000)
+
+
+def instrument(
+    event: str,
+    *,
+    role: str | None = None,
+    keys: Sequence[str] = _STANDARD_KEYS,
+    result_keys: Sequence[str] = (),
+    **static: Any,
+) -> Callable[[Callable[..., Awaitable[T]]], Callable[..., Awaitable[T]]]:
+    """Decorate an async platform op so it emits a structured event per call.
+
+    Fields are pulled from the call's bound arguments (``keys`` present on the
+    signature), plus a static ``role`` and any ``static`` attributes, plus
+    ``duration_ms`` and an ``outcome`` derived from the result (or ``error`` on
+    exception). ``result_keys`` lifts named keys out of a dict return value onto
+    the event (e.g. the ``cancelled`` count from the reaper).
+
+    The emit is wrapped so a telemetry bug can never fail the underlying call.
+    """
+
+    def deco(fn: Callable[..., Awaitable[T]]) -> Callable[..., Awaitable[T]]:
+        sig = inspect.signature(fn)
+
+        @functools.wraps(fn)
+        async def wrapper(*args: Any, **kwargs: Any) -> T:
+            try:
+                bound = sig.bind(*args, **kwargs)
+                bound.apply_defaults()
+                arguments = bound.arguments
+            except TypeError:
+                arguments = {}
+            fields: dict[str, Any] = {k: arguments.get(k) for k in keys}
+            if role is not None:
+                fields["agent_role"] = role
+            fields.update(static)
+
+            t0 = time.monotonic()
+            try:
+                result = await fn(*args, **kwargs)
+            except Exception as exc:
+                emit(event, outcome="error", error=repr(exc), duration_ms=_duration_ms(t0), **fields)
+                raise
+            if isinstance(result, dict):
+                for rk in result_keys:
+                    if rk in result:
+                        fields[rk] = result[rk]
+            emit(event, outcome=_outcome_of(result), duration_ms=_duration_ms(t0), **fields)
+            return result
+
+        return wrapper
+
+    return deco

web/app.py

@@ -57,28 +57,42 @@ from web.agent_platform_port import (
     get_project_url,
     pc_accept_interaction,
     pc_archive_project,
+    pc_cancel_interaction,
     pc_cancel_run,
     pc_create_project,
     pc_create_workflow_issue,
+    pc_escalate_issue,
+    pc_get_agent_health,
+    pc_get_recent_escalations,
+    pc_get_predecessor_context,
+    pc_get_predecessor_for_case,
     pc_get_agents,
     pc_get_agents_for_case,
     pc_get_case_issues,
+    pc_pick_default_comment_target,
     pc_get_issue_comments,
     pc_get_issue_interactions,
     pc_get_run_events,
     pc_get_run_log,
     pc_list_live_runs,
     pc_post_comment,
+    pc_reap_stale_interactions,
     pc_reject_interaction,
     pc_request,
     pc_reset_agent_session,
     pc_reset_case_agents,
     pc_respond_to_interaction,
+    pc_set_issue_status,
     pc_restore_project,
     pc_wake_analyst_for_appraiser_facts,
     pc_wake_analyst_for_argument_aggregation,
+    pc_wake_analyst_for_protocol_analysis,
+    pc_get_generation_run_status,
     rename_case_project,
     pc_wake_ceo,
+    pc_wake_ceo_for_action,
+    pc_open_ceo_run,
+    pc_mark_comment_routed,
     pc_wake_ceo_for_feedback_fold,
     pc_wake_curator_for_final,
     pc_wake_for_precedent_extraction,
@@ -100,17 +114,84 @@ PROGRESS_TTL_SECONDS = 300
 _progress = ProgressStore(config.REDIS_URL, ttl_seconds=PROGRESS_TTL_SECONDS)
 
 
+async def _interaction_reaper_loop():
+    """Safety-net reaper for stranded pending interactions (TaskMaster #215).
+
+    Every 15 min, cancel pending interactions on closed (done/cancelled) issues
+    so the chair's "awaiting you" count never accumulates dead questions.
+    """
+    while True:
+        try:
+            await asyncio.sleep(900)
+            await pc_reap_stale_interactions()
+        except asyncio.CancelledError:
+            raise
+        except Exception:
+            logger.warning("interaction reaper sweep failed", exc_info=True)
+
+
+# Watchdog kill-switch — set AGENT_ZOMBIE_AUTO_ESCALATE=0 to observe health
+# without acting (escalation stays available manually via the endpoint).
+_ZOMBIE_AUTO_ESCALATE = os.environ.get("AGENT_ZOMBIE_AUTO_ESCALATE", "1") != "0"
+
+
+async def _zombie_escalation_loop():
+    """Watchdog: auto-escalate persistent recovery-loop zombies to the chair (#218+#222).
+
+    Every 15 min, derive per-issue health (#222) and hand any persistent zombie
+    (a recovery loop, ``>=`` ``ZOMBIE_ESCALATE_MIN_RECOVERY`` recovery wakeups)
+    to the human via the loop-safe escalate primitive (#218) — closing the loop
+    that today burns budget until a human notices. Server-side on purpose: an
+    agent stuck in a recovery loop cannot reliably self-escalate. Idempotent —
+    an escalated issue is agent-unowned next sweep, so it is not re-escalated.
+    """
+    from web.agent_health import zombie_escalation
+
+    while True:
+        try:
+            await asyncio.sleep(900)
+            if not _ZOMBIE_AUTO_ESCALATE:
+                continue
+            health = await pc_get_agent_health()
+            for item in health.get("items", []):
+                decision = zombie_escalation(item)
+                if decision is None:
+                    continue
+                severity, reason = decision
+                result = await pc_escalate_issue(item["issue_id"], severity, reason)
+                if result.get("ok"):
+                    logger.warning(
+                        "watchdog auto-escalated %s (%s, %d recovery wakeups) to chair [%s]",
+                        item.get("identifier"), item.get("agent_name"),
+                        item.get("recovery_wakeups", 0), severity,
+                    )
+                else:
+                    logger.warning(
+                        "watchdog escalation failed for %s: %s",
+                        item.get("identifier"), result.get("error"),
+                    )
+        except asyncio.CancelledError:
+            raise
+        except Exception:
+            logger.warning("zombie escalation sweep failed", exc_info=True)
+
+
 @asynccontextmanager
 async def lifespan(app: FastAPI):
     UPLOAD_DIR.mkdir(parents=True, exist_ok=True)
     await db.init_schema()
     sync_task = asyncio.create_task(git_sync.sweep_loop())
+    reaper_task = asyncio.create_task(_interaction_reaper_loop())
+    watchdog_task = asyncio.create_task(_zombie_escalation_loop())
     try:
         yield
     finally:
         sync_task.cancel()
+        reaper_task.cancel()
+        watchdog_task.cancel()
+        for _t in (sync_task, reaper_task, watchdog_task):
             try:
-            await sync_task
+                await _t
             except asyncio.CancelledError:
                 pass
         await db.close_pool()
@@ -1828,6 +1909,16 @@ async def health():
     return {"status": "ok"}
 
 
+@app.get("/api/status-model")
+async def api_status_model():
+    """Canonical case-status model — ordered statuses (key/label/description/
+    phase/selectable/terminal/on_enter) + the 5 phases. Single source of truth
+    (legal_mcp/case_status_model.py); the frontend case-status.ts mirrors it."""
+    from legal_mcp.case_status_model import to_dict
+
+    return to_dict()
+
+
 @app.get("/api/cases")
 async def list_cases(
     detail: bool = False,
@@ -2627,6 +2718,97 @@ async def api_get_legal_arguments(case_number: str, party: str = ""):
     }
 
 
+@app.get("/api/cases/{case_number}/protocol-analysis")
+async def api_get_protocol_analysis(case_number: str):
+    """Hearing-protocol comparative analysis for the "מה קרה בדיון" panel (#226).
+
+    Read-only surface over the case-knowledge produced by ``analyze_protocol``:
+    the hearing header (date + who appeared, from the canonical case columns)
+    plus the per-argument verdicts (strengthened / newly_raised / dropped)
+    grouped by party. Container-safe — pure DB reads, no LLM. Returns an empty
+    analysis (not an error) when the analysis has not been run yet.
+    """
+    case = await db.get_case_by_number(case_number)
+    if not case:
+        raise HTTPException(404, f"תיק {case_number} לא נמצא")
+
+    case_id = UUID(case["id"])
+    rows = await db.list_protocol_analysis(case_id)
+
+    # Protocol title comes from the analysed document (all rows share one
+    # document_id per idempotent replace). Resolve it via the case's documents.
+    protocol_title = ""
+    protocol_document_id = ""
+    if rows:
+        protocol_document_id = rows[0].get("document_id") or ""
+        docs = await db.list_documents(case_id)
+        match = next(
+            (d for d in docs if str(d.get("id")) == protocol_document_id), None,
+        )
+        if match:
+            protocol_title = match.get("title") or ""
+
+    attendees = case.get("hearing_attendees") or {}
+    hearing_date = case.get("hearing_date")
+    header = {
+        "hearing_date": hearing_date.isoformat() if hearing_date else None,
+        "protocol_title": protocol_title,
+        "protocol_document_id": protocol_document_id,
+        "panel_members": attendees.get("panel_members") or [],
+        "appellants_present": attendees.get("appellants_present") or [],
+        "respondents_present": attendees.get("respondents_present") or [],
+    }
+
+    # Group verdicts by party for the panel's per-side sections, in display order.
+    by_party: dict[str, list[dict]] = {}
+    by_change: dict[str, int] = {}
+    for r in rows:
+        by_party.setdefault(r.get("party_role") or "", []).append(r)
+        ct = r.get("change_type", "")
+        by_change[ct] = by_change.get(ct, 0) + 1
+
+    return {
+        "case_number": case_number,
+        "total": len(rows),
+        "header": header,
+        "by_change": by_change,
+        "by_party": by_party,
+        "analysis": rows,
+    }
+
+
+@app.post("/api/cases/{case_number}/analyze-protocol")
+async def api_analyze_protocol(case_number: str, document_id: str = ""):
+    """Queue hearing-protocol analysis by waking the legal-analyst agent (#226).
+
+    Same delegation rationale as ``aggregate-arguments``: ``analyze_protocol``
+    calls the local ``claude`` CLI, absent in this container, so we route to the
+    company's analyst rather than running a doomed in-container BackgroundTask.
+
+    Response: {"status": "queued", ...} or {"status": "skipped", "reason": ...}.
+    """
+    case = await db.get_case_by_number(case_number)
+    if not case:
+        raise HTTPException(404, f"תיק {case_number} לא נמצא")
+
+    prefix = case_number[:1]
+    company_id = (
+        PAPERCLIP_COMPANIES["licensing"] if prefix == "1"
+        else PAPERCLIP_COMPANIES["betterment"] if prefix in ("8", "9")
+        else ""
+    )
+
+    try:
+        result = await pc_wake_analyst_for_protocol_analysis(
+            case_number, company_id=company_id, document_id=document_id,
+        )
+    except Exception as e:
+        logger.exception("analyst wakeup failed for protocol analysis %s", case_number)
+        raise HTTPException(500, f"לא ניתן לשלוח לאנליטיקאי: {e}")
+
+    return result
+
+
 @app.post("/api/cases/{case_number}/direction")
 async def api_set_direction(case_number: str, req: DirectionRequest):
     """Save the approved direction document for the discussion block."""
@@ -2889,42 +3071,82 @@ async def api_learn(case_number: str):
 # ── Local files API — research, drafts, proofread ──
 
 
+# Directories under a case that are infrastructure/cache, not chair-facing
+# content. Hidden dirs (``.git``, ``.claude`` …) are skipped separately.
+_CASE_FILE_EXCLUDE_DIRS = {"thumbnails"}
+
+
+def _is_content_dir(p: Path) -> bool:
+    return p.is_dir() and not p.name.startswith(".") and p.name not in _CASE_FILE_EXCLUDE_DIRS
+
+
+def _case_file_folders(case_dir: Path) -> list[tuple[str, Path]]:
+    """Enumerate every content subdirectory under a case dir as (key, path).
+
+    Covers *all* subfolders (not just research) — recursing one level into
+    ``documents/`` — while skipping hidden (``.git``/``.claude``…) and cache
+    dirs. The ``key`` is the case-relative path with ``/`` encoded as ``__`` so
+    it survives a single URL path segment; ``_resolve_case_file`` re-validates it
+    against this same enumeration, so the key set is the serve allowlist (no path
+    traversal).
+    """
+    folders: list[tuple[str, Path]] = []
+    if not case_dir.exists():
+        return folders
+    for child in sorted(case_dir.iterdir()):
+        if not _is_content_dir(child):
+            continue
+        if child.name == "documents":
+            folders.extend(
+                (f"documents__{sub.name}", sub)
+                for sub in sorted(child.iterdir())
+                if _is_content_dir(sub)
+            )
+        else:
+            folders.append((child.name, child))
+    return folders
+
+
+def _resolve_case_file(case_dir: Path, folder_key: str, filename: str) -> Path | None:
+    """Safely resolve (folder_key, filename) to a file inside the case dir.
+
+    Returns None unless ``folder_key`` is one of the enumerated content folders
+    AND ``filename`` resolves to a regular file directly inside it (no traversal).
+    """
+    base = dict(_case_file_folders(case_dir)).get(folder_key)
+    if base is None:
+        return None
+    target = (base / filename).resolve()
+    if target.parent != base.resolve() or not target.is_file():
+        return None
+    return target
+
+
 @app.get("/api/cases/{case_number}/local-files")
 async def api_local_files(case_number: str):
-    """List local files from case subdirectories (research, drafts, proofread)."""
+    """List local files across ALL content subfolders under a case (not just
+    research) — so the chair can reach any agent-produced file (drafts, OCR text,
+    originals…) from the UI instead of only over SSH. Empty folders are returned
+    too (with an empty ``files`` list) so the UI can show them as such."""
     case_dir = config.find_case_dir(case_number)
-    result = {}
-    for folder in ("research", "proofread"):
-        folder_path = case_dir / "documents" / folder
-        if folder_path.exists():
+    folders = []
+    for key, path in _case_file_folders(case_dir):
         files = []
-            for f in sorted(folder_path.iterdir()):
+        for f in sorted(path.iterdir()):
             if f.is_file() and not f.name.startswith("."):
                 stat = f.stat()
                 files.append({
                     "filename": f.name,
                     "size": stat.st_size,
                     "modified_at": stat.st_mtime,
-                        "folder": folder,
                 })
-            if files:
-                result[folder] = files
-    # Drafts are at case level, not under documents
-    drafts_path = case_dir / "drafts"
-    if drafts_path.exists():
-        files = []
-        for f in sorted(drafts_path.iterdir()):
-            if f.is_file() and not f.name.startswith("."):
-                stat = f.stat()
-                files.append({
-                    "filename": f.name,
-                    "size": stat.st_size,
-                    "modified_at": stat.st_mtime,
-                    "folder": "drafts",
+        folders.append({
+            "key": key,
+            "name": path.name,
+            "path": str(path.relative_to(case_dir)),
+            "files": files,
         })
-        if files:
-            result["drafts"] = files
-    return result
+    return {"folders": folders}
 
 
 async def serve_blob(
@@ -2994,17 +3216,32 @@ async def _seal_blob_file(dest: Path, *, bucket=storage.Bucket.DOCUMENTS) -> Non
 
 @app.get("/api/cases/{case_number}/local-files/{folder}/{filename}")
 async def api_read_local_file(case_number: str, folder: str, filename: str):
-    """Read contents of a local case file."""
-    if folder not in ("research", "proofread", "drafts"):
-        raise HTTPException(400, "Invalid folder")
+    """Serve a single local case file from any content subfolder.
+
+    ``folder`` is a key from ``/local-files`` (``documents__research``, ``drafts``…);
+    ``_resolve_case_file`` validates it against the folder allowlist and blocks
+    path traversal before the bytes are served.
+    """
     case_dir = config.find_case_dir(case_number)
-    if folder == "drafts":
-        path = case_dir / "drafts" / filename
-    else:
-        path = case_dir / "documents" / folder / filename
-    if not path.exists() or not path.is_file():
+    path = _resolve_case_file(case_dir, folder, filename)
+    if path is None:
         raise HTTPException(404, "קובץ לא נמצא")
-    return await serve_blob(path, media_type="text/plain; charset=utf-8", filename=filename)
+    media_type = _local_file_media_type(filename)
+    return await serve_blob(path, media_type=media_type, filename=filename)
+
+
+def _local_file_media_type(filename: str) -> str:
+    """Best-effort content type for a served case file (text renders inline,
+    binaries download)."""
+    ext = filename.rsplit(".", 1)[-1].lower() if "." in filename else ""
+    return {
+        "md": "text/markdown; charset=utf-8",
+        "txt": "text/plain; charset=utf-8",
+        "json": "application/json; charset=utf-8",
+        "pdf": "application/pdf",
+        "docx": "application/vnd.openxmlformats-officedocument.wordprocessingml.document",
+        "doc": "application/msword",
+    }.get(ext, "application/octet-stream")
 
 
 # ── Research analysis (analysis-and-research.md) — parse + edit ────
@@ -3080,6 +3317,140 @@ async def api_party_claims_summary(case_number: str):
     return {"markdown": path.read_text(encoding="utf-8")}
 
 
+# ── Deterministic document generation triggers (TaskMaster #214) ─────
+# The "הפקת מסמכים" buttons in the case "טיוטות והערות" tab. Generation for the
+# executive summary + interim draft runs HOST-SIDE (claude_session → claude -p,
+# not in the container), so these endpoints fire a *structured-action* CEO wakeup
+# (the CEO reads payload.action and routes to שלב H2 / שלב H — legal-ceo.md).
+# Wakeup goes through the Paperclip API helper (pc_wake_ceo_for_action → the
+# platform Port → POST /api/agents/{id}/wakeup), NEVER a direct DB insert.
+# Polling for completion uses the existing read endpoints (research/party-claims-
+# summary for the summary; the exports list for the טיוטה-טענות_הצדדים_*.docx file).
+
+
+async def _wake_ceo_action(case_number: str, action: str) -> dict:
+    """Resolve the case's company + fire a structured-action CEO wakeup (#214)."""
+    case = await db.get_case_by_number(case_number)
+    if not case:
+        raise HTTPException(404, f"תיק {case_number} לא נמצא")
+    prefix = case_number[:1]
+    company_id = (
+        PAPERCLIP_COMPANIES["licensing"] if prefix == "1"
+        else PAPERCLIP_COMPANIES["betterment"] if prefix in ("8", "9")
+        else ""
+    )
+    try:
+        return await pc_wake_ceo_for_action(case_number, action, company_id=company_id)
+    except Exception as e:
+        logger.warning("CEO action %s wakeup failed for %s: %s", action, case_number, e)
+        return {"status": "error", "error": str(e)}
+
+
+@app.post("/api/cases/{case_number}/generate/party-claims-summary", status_code=202)
+async def api_generate_party_claims_summary(case_number: str):
+    """Fire-and-accept: wake the CEO to generate the party-claims executive summary
+    (שלב H2 → summarize_party_claims). Runs host-side; poll
+    GET /api/cases/{n}/research/party-claims-summary for completion."""
+    return await _wake_ceo_action(case_number, "party_claims_summary")
+
+
+@app.post("/api/cases/{case_number}/generate/interim-draft", status_code=202)
+async def api_generate_interim_draft(case_number: str):
+    """Fire-and-accept: wake the CEO to generate the partial "party-claims" draft
+    (שלב H → write_interim_draft + export_interim_draft). Runs host-side; poll the
+    exports list for the new טיוטה-טענות_הצדדים_{N}.docx."""
+    return await _wake_ceo_action(case_number, "interim_draft")
+
+
+_GENERATE_ACTIONS = {"party_claims_summary", "interim_draft"}
+# Run statuses (Paperclip heartbeat_runs.status) grouped for state mapping.
+_RUN_ACTIVE = {"running", "in_progress", "queued", "pending", "scheduled", "created"}
+_RUN_FAILED = {"cancelled", "failed", "error", "timed_out", "lost"}
+
+
+def _generation_output_ready(case_number: str, action: str) -> bool:
+    """Whether the generation artifact already exists on disk (host-side)."""
+    if action == "party_claims_summary":
+        from legal_mcp.services import party_claims_summary as pcs
+        return pcs.summary_file_path(case_number).exists()
+    # interim_draft → a "טיוטה-טענות_הצדדים_*.docx" file in the case exports dir.
+    export_dir = config.find_case_dir(case_number) / "exports"
+    if not export_dir.exists():
+        return False
+    return any(
+        f.is_file() and f.suffix.lower() == ".docx" and f.name.startswith("טיוטה-טענות_הצדדים_")
+        for f in export_dir.iterdir()
+    )
+
+
+@app.get("/api/cases/{case_number}/generate/{action}/status")
+async def api_generate_status(case_number: str, action: str):
+    """Server-derived status of a "הפקת מסמכים" background generation (#227 ג).
+
+    Reconstructs {idle|queued|running|done|failed} from the CEO child issue + its
+    latest heartbeat_run (via the platform port) plus whether the output artifact
+    exists — so the UI indicator is correct after navigating away and back; it
+    never depends on the browser staying on the page. ``started_at``/``finished_at``
+    are server timestamps (ISO-8601 UTC) so the client renders a correct elapsed
+    time on return.
+    """
+    if action not in _GENERATE_ACTIONS:
+        raise HTTPException(400, f"action לא תקין: {action}")
+    case = await db.get_case_by_number(case_number)
+    if not case:
+        raise HTTPException(404, f"תיק {case_number} לא נמצא")
+
+    output_ready = _generation_output_ready(case_number, action)
+    try:
+        run = await pc_get_generation_run_status(case_number, action)
+    except Exception as e:
+        logger.warning("generation status lookup failed for %s/%s: %s", case_number, action, e)
+        run = {"issue_id": None}
+
+    run_status = (run.get("run_status") or "").lower()
+    started_at = run.get("started_at")
+    active = run_status in _RUN_ACTIVE or (
+        run.get("issue_status") == "in_progress"
+        and run_status not in _RUN_FAILED
+        and run_status not in {"completed", "succeeded", "done"}
+        and not run.get("finished_at")
+    )
+
+    # Precedence: an in-flight run wins; then a produced artifact; then a failed
+    # run; else idle. (A produced file reads as "done" even if a later retry
+    # failed — the artifact exists.)
+    if active:
+        state = "running" if started_at else "queued"
+    elif output_ready:
+        state = "done"
+    elif run_status in _RUN_FAILED:
+        state = "failed"
+    elif run_status in {"completed", "succeeded", "done"}:
+        # Ran to completion but produced no artifact → treat as a failure so the
+        # chair isn't left thinking it worked.
+        state = "failed"
+    else:
+        state = "idle"
+
+    detail = ""
+    if state == "failed":
+        if run.get("error_code") == "issue_assignee_changed":
+            detail = "הריצה בוטלה (הוקצתה מחדש)"
+        else:
+            detail = run.get("error") or "הריצה נכשלה"
+
+    return {
+        "case_number": case_number,
+        "action": action,
+        "state": state,
+        "started_at": started_at,
+        "finished_at": run.get("finished_at"),
+        "output_ready": output_ready,
+        "detail": detail,
+        "issue_id": run.get("issue_id"),
+    }
+
+
 @app.get("/api/cases/{case_number}/research/party-claims-summary/download")
 async def api_party_claims_summary_download(case_number: str):
     """Download the raw party-claims-summary.md file."""
@@ -4291,14 +4662,39 @@ async def api_case_agents(case_number: str):
 class AgentCommentRequest(BaseModel):
     body: str
     issue_id: str | None = None
+    # "פתח הרצה חדשה" — skip the comment entirely and only open a CEO run. Kept for
+    # the agents-tab target selector; the default path below now opens a run too, so
+    # this only differs in *not* recording the instruction on the parent thread.
+    new_run: bool = False
 
 
 @app.post("/api/cases/{case_number}/agents/comment")
 async def api_post_agent_comment(case_number: str, req: AgentCommentRequest):
-    """Post a comment on a Paperclip issue linked to a case.
+    """Post a chair instruction to a Paperclip issue linked to a case.
 
-    If issue_id is omitted, the most recent non-done issue is used.
+    Routing (agents-tab target selector):
+      - ``new_run=True`` → open a fresh CEO-owned run (child issue assigned to the
+        CEO + wakeup), sidestepping the ``issue_assignee_changed`` cancellation on
+        a human-owned parent. Overrides ``issue_id``.
+      - ``issue_id`` given → post to that issue (explicit chair choice).
+      - neither → prefer the live CEO main issue, never a closed one
+        (``pick_default_comment_target``).
     """
+    if req.new_run:
+        result = await pc_open_ceo_run(case_number, req.body)
+        if result.get("status") != "ok":
+            raise HTTPException(
+                502,
+                f"פתיחת הרצה חדשה נכשלה: {result.get('reason', 'unknown')}",
+            )
+        return {
+            "comment_id": None,
+            "issue_id": result["issue_id"],
+            "issue_identifier": result.get("identifier", ""),
+            "issue_status": "in_progress",
+            "new_run": True,
+        }
+
     issues = await pc_get_case_issues(case_number)
     if not issues:
         raise HTTPException(404, f"לא נמצא פרויקט Paperclip לתיק {case_number}")
@@ -4308,14 +4704,35 @@ async def api_post_agent_comment(case_number: str, req: AgentCommentRequest):
         if not target:
             raise HTTPException(404, f"Issue {req.issue_id} לא שייך לתיק {case_number}")
     else:
-        # Pick the most recent non-done issue, or the last one
-        active = [i for i in issues if i["status"] != "done"]
-        target = active[-1] if active else issues[-1]
+        target = pc_pick_default_comment_target(issues)
 
     result = await pc_post_comment(target["id"], target["company_id"], req.body)
 
-    # Find the identifier for the response
+    # The comment alone reaches nobody: a case issue awaiting the chair is
+    # `in_review` and owned by a human, so Paperclip cancels any wakeup on it
+    # (`issue_assignee_changed` → "wake the new owner" → the owner is a person).
+    # The plugin sweep that was meant to catch this delivers the instruction as
+    # `payload.prompt`, which the runner drops. So the comment is recorded for the
+    # thread, and the instruction is carried by a CEO-owned child run — the one
+    # path verified to reach the agent (TaskMaster #228; see paperclip-quirks.md).
+    run = await pc_open_ceo_run(case_number, req.body)
+    if run.get("status") != "ok":
+        raise HTTPException(
+            502,
+            f"ההערה נשמרה אך פתיחת ההרצה נכשלה — הסוכן לא יטפל בה: "
+            f"{run.get('reason', 'unknown')}",
+        )
+    # Claim the comment so the sweep does not re-route it: the CEO answers on the
+    # child issue, leaving the parent with no agent reply, which the sweep reads
+    # as "still pending" forever.
+    await pc_mark_comment_routed(target["id"], result.get("comment_id", ""))
+
+    # Echo the resolved target so the UI can show where it landed and flag it
+    # when the chair explicitly targeted a closed issue (whose wakeup is skipped).
     result["issue_identifier"] = target.get("identifier", "")
+    result["issue_status"] = target.get("status", "")
+    result["run_issue_id"] = run["issue_id"]
+    result["run_issue_identifier"] = run.get("identifier", "")
     return result
 
 
@@ -4332,8 +4749,10 @@ async def api_post_interaction_response(
 ):
     """Submit a user's answer to a Paperclip issue-thread interaction.
 
-    Routes to /respond | /accept | /reject based on `action`. Paperclip
-    auto-wakes the issue assignee after a successful submission.
+    Routes to /respond | /accept | /reject based on `action`, then opens a CEO run
+    carrying the answer. Paperclip's own `wake_assignee` cannot deliver it: the
+    issue is owned by the chair while it waits for her, so the "assignee" it wakes
+    is a person and no agent ever runs (TaskMaster #228).
     """
     issues = await pc_get_case_issues(case_number)
     if not any(i["id"] == req.issue_id for i in issues):
@@ -4345,7 +4764,7 @@ async def api_post_interaction_response(
         "reject": pc_reject_interaction,
     }
     try:
-        return await handlers[req.action](
+        result = await handlers[req.action](
             req.issue_id, req.interaction_id, req.payload,
         )
     except httpx.HTTPStatusError as e:
@@ -4354,6 +4773,47 @@ async def api_post_interaction_response(
     except Exception as e:
         raise HTTPException(502, f"שגיאת Paperclip: {e}")
 
+    answer = json.dumps(req.payload, ensure_ascii=False)
+    run = await pc_open_ceo_run(
+        case_number,
+        f'תשובת היו"ר לשאלה שהוצגה לה ({req.action}):\n\n{answer}\n\n'
+        f"קרא את ה-interaction המקורי על ה-issue והמשך משם.",
+    )
+    if run.get("status") != "ok":
+        raise HTTPException(
+            502,
+            f"התשובה נשמרה אך פתיחת ההרצה נכשלה — הסוכן לא ימשיך: "
+            f"{run.get('reason', 'unknown')}",
+        )
+    if isinstance(result, dict):
+        result["run_issue_id"] = run["issue_id"]
+        result["run_issue_identifier"] = run.get("identifier", "")
+    return result
+
+
+class InteractionDismissRequest(BaseModel):
+    issue_id: str
+    interaction_id: str
+
+
+@app.post("/api/cases/{case_number}/agents/interaction-dismiss")
+async def api_dismiss_interaction(
+    case_number: str, req: InteractionDismissRequest,
+):
+    """Dismiss a pending interaction WITHOUT waking the issue assignee.
+
+    For stale/duplicate questions (TaskMaster #215). Unlike interaction-response,
+    this does not resolve via Paperclip's wake_assignee path — it cancels the row
+    so the chair can clear noise without triggering an agent run.
+    """
+    issues = await pc_get_case_issues(case_number)
+    if not any(i["id"] == req.issue_id for i in issues):
+        raise HTTPException(404, f"Issue {req.issue_id} לא שייך לתיק {case_number}")
+    result = await pc_cancel_interaction(req.issue_id, req.interaction_id)
+    if not result.get("ok"):
+        raise HTTPException(404, result.get("error", "ביטול הבקשה נכשל"))
+    return result
+
 
 @app.post("/api/cases/{case_number}/agents/reset")
 async def api_reset_case_agents(case_number: str):
@@ -4368,6 +4828,53 @@ async def api_reset_case_agents(case_number: str):
     return result
 
 
+class EscalateRequest(BaseModel):
+    issue_id: str
+    severity: Literal["critical", "high", "medium"]
+    reason: str
+
+
+@app.post("/api/cases/{case_number}/agents/escalate")
+async def api_escalate_issue(case_number: str, req: EscalateRequest):
+    """Escalate a single stuck issue to the chair (#218).
+
+    The loop-safe alternative to leaving an issue agent-owned+blocked: hands the
+    issue to the human in one atomic transition (in_review + assignee_user_id)
+    and records a severity note, without re-waking any agent. Emits an
+    ``agent.escalated`` telemetry event via the Port.
+    """
+    issues = await pc_get_case_issues(case_number)
+    if not any(i["id"] == req.issue_id for i in issues):
+        raise HTTPException(404, f"Issue {req.issue_id} לא שייך לתיק {case_number}")
+    result = await pc_escalate_issue(req.issue_id, req.severity, req.reason)
+    if not result.get("ok"):
+        raise HTTPException(400, result.get("error", "ההסלמה נכשלה"))
+    return result
+
+
+class IssueStatusRequest(BaseModel):
+    issue_id: str
+    status: Literal["done", "cancelled"]
+
+
+@app.post("/api/cases/{case_number}/agents/issue-status")
+async def api_set_issue_status(case_number: str, req: IssueStatusRequest):
+    """Manually close a board issue to done/cancelled from the chair UI.
+
+    Lets the chair tidy the agents board by hand — superseded "הרצה חדשה" runs
+    left in ``in_review`` had no manual close control and piled up. Only the two
+    terminal statuses are accepted; the loop-safe direct-DB close lives behind the
+    Port (``pc_set_issue_status``) and issues no wakeup.
+    """
+    issues = await pc_get_case_issues(case_number)
+    if not any(i["id"] == req.issue_id for i in issues):
+        raise HTTPException(404, f"Issue {req.issue_id} לא שייך לתיק {case_number}")
+    result = await pc_set_issue_status(req.issue_id, req.status)
+    if not result.get("ok"):
+        raise HTTPException(400, result.get("error", "עדכון הסטטוס נכשל"))
+    return result
+
+
 # ── Settings: MCP Server Configuration ────────────────────────────
 #
 # Source of truth for legal-ai env vars is Coolify (see memory:
@@ -5621,13 +6128,21 @@ async def api_reprocess_document(case_number: str, doc_id: str):
 
 
 _ALLOWED_APPRAISER_SIDES = {"committee", "appellant", "deciding"}
+# metadata.protocol_scope — only meaningful when doc_type == "protocol".
+# "appeal" = ועדת הערר (default when absent), "lower" = ועדה מקומית/מחוזית.
+# Used by the overview panel to sort a protocol into group 3 vs group 2.
+_ALLOWED_PROTOCOL_SCOPES = {"appeal", "lower"}
 
 
 class DocumentPatchRequest(BaseModel):
-    """Patch payload for a single document. Both fields are optional."""
+    """Patch payload for a single document. All fields are optional; only the
+    ones present are applied. The metadata.* flags are display/processing hints
+    that never change the stored doc_type."""
 
     doc_type: str | None = None
     appraiser_side: str | None = None  # committee | appellant | deciding | "" to clear
+    is_post_hearing: bool | None = None  # material submitted after the hearing (→ group 4, block-chet)
+    protocol_scope: str | None = None  # appeal | lower | "" to clear (protocol docs only)
 
 
 @app.patch("/api/cases/{case_number}/documents/{doc_id}")
@@ -5652,6 +6167,13 @@ async def api_patch_document(case_number: str, doc_id: str, req: DocumentPatchRe
 
     updates: dict = {}
 
+    # Load metadata once — appraiser_side, is_post_hearing and protocol_scope
+    # all live in the same JSONB blob, so a single patch may touch several.
+    metadata = doc.get("metadata") or {}
+    if isinstance(metadata, str):
+        metadata = json.loads(metadata)
+    metadata_dirty = False
+
     if req.doc_type is not None:
         if req.doc_type not in DOC_TYPE_NAMES:
             raise HTTPException(
@@ -5668,13 +6190,34 @@ async def api_patch_document(case_number: str, doc_id: str, req: DocumentPatchRe
                 f"appraiser_side לא תקין: {req.appraiser_side}. ערכים מותרים: "
                 f"{', '.join(sorted(_ALLOWED_APPRAISER_SIDES))}",
             )
-        metadata = doc.get("metadata") or {}
-        if isinstance(metadata, str):
-            metadata = json.loads(metadata)
         if req.appraiser_side:
             metadata["appraiser_side"] = req.appraiser_side
         else:
             metadata.pop("appraiser_side", None)
+        metadata_dirty = True
+
+    if req.is_post_hearing is not None:
+        # Store True; drop the key entirely when False so absence == "before".
+        if req.is_post_hearing:
+            metadata["is_post_hearing"] = True
+        else:
+            metadata.pop("is_post_hearing", None)
+        metadata_dirty = True
+
+    if req.protocol_scope is not None:
+        if req.protocol_scope and req.protocol_scope not in _ALLOWED_PROTOCOL_SCOPES:
+            raise HTTPException(
+                422,
+                f"protocol_scope לא תקין: {req.protocol_scope}. ערכים מותרים: "
+                f"{', '.join(sorted(_ALLOWED_PROTOCOL_SCOPES))}",
+            )
+        if req.protocol_scope:
+            metadata["protocol_scope"] = req.protocol_scope
+        else:
+            metadata.pop("protocol_scope", None)
+        metadata_dirty = True
+
+    if metadata_dirty:
         updates["metadata"] = metadata
 
     if not updates:
@@ -7187,6 +7730,50 @@ async def operations_agents():
     }
 
 
+@app.get("/api/operations/agents/health")
+async def operations_agent_health():
+    """Per-issue agent health taxonomy (#222): zombie / stalled / working / idle.
+
+    Read-only. Surfaces the stranded-child / recovery-loop cases (``zombie``)
+    automatically — the ``gt feed --problems`` idea grounded in our failure
+    modes — so they no longer need hand-querying the Paperclip DB. Consumed by
+    the /operations dashboard (UI is design-gated, follow-up).
+    """
+    return await pc_get_agent_health()
+
+
+@app.get("/api/operations/agents/escalations")
+async def operations_agent_escalations(limit: int = 10, hours: int = 24):
+    """Recent chair escalations — watchdog + manual (#218/#222).
+
+    Read-only history of what was handed to the chair, for the ops health panel.
+    """
+    return await pc_get_recent_escalations(limit, hours)
+
+
+@app.get("/api/operations/issues/{issue_id}/predecessor")
+async def operations_issue_predecessor(issue_id: str, limit: int = 3):
+    """Recent finished runs on an issue for session continuation (#220, "seance").
+
+    Read-only. Lets a fresh heartbeat read what its predecessors on the same
+    issue concluded (the run ``summary`` each leaves) instead of rediscovering
+    context from scratch. Agent-facing wiring (MCP tool + HEARTBEAT) is a
+    follow-up; this is the backend it will call.
+    """
+    return await pc_get_predecessor_context(issue_id, limit)
+
+
+@app.get("/api/operations/cases/{case_number}/predecessor")
+async def operations_case_predecessor(case_number: str, limit: int = 5):
+    """Recent finished runs across a case's issues (#220, agent-facing "seance").
+
+    Case-scoped sibling of the by-issue endpoint — the plugin tool
+    ``legal_predecessor_context`` calls this so a resuming agent reads what prior
+    sessions on the case concluded (agents key on case_number, not issue UUID).
+    """
+    return await pc_get_predecessor_for_case(case_number, limit)
+
+
 @app.get("/api/operations/agents/runs/{run_id}/log")
 async def operations_agent_run_log(run_id: str):
     """Full output log (NDJSON stream) of one heartbeat run."""