Add drafts & feedback tab to case page, remove global feedback page

Move draft management (export DOCX, download, upload revised version, mark
final) and chair feedback into a new "טיוטות והערות" tab on the case detail
page. Remove the standalone /feedback page and its nav link since feedback
is now case-scoped.

Also fix /api/admin/skills 500 error when Paperclip DB is unreachable by
adding a connection timeout and graceful fallback to disk-only skills.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

web/app.py

@@ -1102,7 +1102,23 @@ async def api_case_create(req: CaseCreateRequest):
         practice_area=req.practice_area,
         appeal_subtype=req.appeal_subtype,
     )
-    return json.loads(result)
+    parsed = json.loads(result)
+
+    # Auto-create Paperclip project for the new case
+    appeal_type = req.appeal_subtype or "רישוי"
+    try:
+        pc_result = await pc_create_project(
+            case_number=req.case_number,
+            title=req.title,
+            appeal_type=appeal_type,
+        )
+        parsed["paperclip"] = pc_result
+        logger.info("Auto-created Paperclip project for case %s: %s", req.case_number, pc_result.get("url"))
+    except Exception as e:
+        logger.warning("Failed to auto-create Paperclip project for case %s: %s", req.case_number, e)
+        parsed["paperclip_error"] = str(e)
+
+    return parsed
 
 
 @app.get("/api/cases/{case_number}/details")
@@ -2091,15 +2107,20 @@ SKILLS_COMPANY_ID = os.environ.get("PAPERCLIP_COMPANY_ID", "42a7acd0-30c5-4cbd-a
 @app.get("/api/admin/skills")
 async def api_list_skills():
     """List installed Paperclip skills with DB sync status."""
-    conn = await asyncpg.connect(PAPERCLIP_DB_URL)
+    rows = []
     try:
-        rows = await conn.fetch(
-            "SELECT slug, name, length(markdown) as md_chars, file_inventory, updated_at "
-            "FROM company_skills WHERE company_id = $1::uuid ORDER BY slug",
-            SKILLS_COMPANY_ID,
-        )
-    finally:
-        await conn.close()
+        conn = await asyncpg.connect(PAPERCLIP_DB_URL, timeout=5)
+        try:
+            rows = await conn.fetch(
+                "SELECT slug, name, length(markdown) as md_chars, file_inventory, updated_at "
+                "FROM company_skills WHERE company_id = $1::uuid ORDER BY slug",
+                SKILLS_COMPANY_ID,
+            )
+        finally:
+            await conn.close()
+    except (OSError, asyncpg.PostgresError, asyncpg.InterfaceError, TimeoutError):
+        # Paperclip DB unreachable — continue with disk-only skills
+        pass
 
     skills = []
     for r in rows: