feat: complete issue fixes and implement remaining features

## Critical Issues (CRIT-001~003) - All Fixed
- JWT secret key validation with pydantic field_validator
- Login audit logging for success/failure attempts
- Frontend API path prefix removal

## High Priority Issues (HIGH-001~008) - All Fixed
- Project soft delete using is_active flag
- Redis session token bytes handling
- Rate limiting with slowapi (5 req/min for login)
- Attachment API permission checks
- Kanban view with drag-and-drop
- Workload heatmap UI (WorkloadPage, WorkloadHeatmap)
- TaskDetailModal integrating Comments/Attachments
- UserSelect component for task assignment

## Medium Priority Issues (MED-001~012) - All Fixed
- MED-001~005: DB commits, N+1 queries, datetime, error format, blocker flag
- MED-006: Project health dashboard (HealthService, ProjectHealthPage)
- MED-007: Capacity update API (PUT /api/users/{id}/capacity)
- MED-008: Schedule triggers (cron parsing, deadline reminders)
- MED-009: Watermark feature (image/PDF watermarking)
- MED-010~012: useEffect deps, DOM operations, PDF export

## New Files
- backend/app/api/health/ - Project health API
- backend/app/services/health_service.py
- backend/app/services/trigger_scheduler.py
- backend/app/services/watermark_service.py
- backend/app/core/rate_limiter.py
- frontend/src/pages/ProjectHealthPage.tsx
- frontend/src/components/ProjectHealthCard.tsx
- frontend/src/components/KanbanBoard.tsx
- frontend/src/components/WorkloadHeatmap.tsx

## Tests
- 113 new tests passing (health: 32, users: 14, triggers: 35, watermark: 32)

## OpenSpec Archives
- add-project-health-dashboard
- add-capacity-update-api
- add-schedule-triggers
- add-watermark-feature
- add-rate-limiting
- enhance-frontend-ux
- add-resource-management-ui

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

backend/app/services/workload_service.py

@@ -184,12 +184,17 @@ def get_workload_heatmap(
     Returns:
         List of UserWorkloadSummary objects
     """
+    from datetime import datetime
+    from collections import defaultdict
+
     if week_start is None:
         week_start = get_current_week_start()
     else:
         # Normalize to week start (Monday)
         week_start = get_week_bounds(week_start)[0]
 
+    week_start, week_end = get_week_bounds(week_start)
+
     # Build user query
     query = db.query(User).filter(User.is_active == True)
 
@@ -201,10 +206,58 @@ def get_workload_heatmap(
 
     users = query.options(joinedload(User.department)).all()
 
-    # Calculate workload for each user
+    if not users:
+        return []
+
+    # Batch query: fetch all tasks for all users in one query
+    user_id_list = [user.id for user in users]
+    week_start_dt = datetime.combine(week_start, datetime.min.time())
+    week_end_dt = datetime.combine(week_end, datetime.max.time())
+
+    all_tasks = (
+        db.query(Task)
+        .join(Task.status, isouter=True)
+        .filter(
+            Task.assignee_id.in_(user_id_list),
+            Task.due_date >= week_start_dt,
+            Task.due_date <= week_end_dt,
+            # Exclude completed tasks
+            (TaskStatus.is_done == False) | (Task.status_id == None)
+        )
+        .all()
+    )
+
+    # Group tasks by assignee_id in memory
+    tasks_by_user: dict = defaultdict(list)
+    for task in all_tasks:
+        tasks_by_user[task.assignee_id].append(task)
+
+    # Calculate workload for each user using pre-fetched tasks
     results = []
     for user in users:
-        summary = calculate_user_workload(db, user, week_start)
+        user_tasks = tasks_by_user.get(user.id, [])
+
+        # Calculate allocated hours from original_estimate
+        allocated_hours = Decimal("0")
+        for task in user_tasks:
+            if task.original_estimate:
+                allocated_hours += task.original_estimate
+
+        capacity_hours = Decimal(str(user.capacity)) if user.capacity else Decimal("40")
+        load_percentage = calculate_load_percentage(allocated_hours, capacity_hours)
+        load_level = determine_load_level(load_percentage)
+
+        summary = UserWorkloadSummary(
+            user_id=user.id,
+            user_name=user.name,
+            department_id=user.department_id,
+            department_name=user.department.name if user.department else None,
+            capacity_hours=capacity_hours,
+            allocated_hours=allocated_hours,
+            load_percentage=load_percentage,
+            load_level=load_level,
+            task_count=len(user_tasks),
+        )
         results.append(summary)
 
     return results