fix: postgres WAL corruption recovery + memory bump + researcher/executor

- Postgres memory: 256M→384M limits, 64M→96M reservations
- Added pg_resetwal -f pre-deploy step to recover from corrupt WAL
  ('PANIC: could not locate a valid checkpoint record' caused by
  force-killed postgres during --force-recreate)
- Added data-checksums initdb arg for future corruption detection
- api→postgres and web→api depends_on: service_healthy→service_started
- Deploy wait loop: fail fast on unhealthy, wait on starting (180s)
- Added researcher/executor to ValidAssignees and frontend dropdowns

backend/Services/TaskService.cs

@@ -12,7 +12,7 @@ public sealed class TaskService(
     IHttpContextAccessor httpContextAccessor) : ITaskService
 {
     private static readonly HashSet<string> ValidAssignees =
-        ["bao", "iris", "programmer", "reviewer", "architekt"];
+        ["bao", "iris", "programmer", "reviewer", "architekt", "researcher", "executor"];
 
     public async Task<IReadOnlyList<WorkTask>> GetAllAsync(CancellationToken ct = default)
         => await taskRepo.GetAllAsync(ct);