fix(worker): claim Running before creating run resources

TaskRunner.RunAsync created the worktree (PrepareRunDirectoryAsync)
before claiming Running via StartRunningAsync. RunNow dispatches by
task id with no atomic claim of their own, so a Queued task racing
the queue picker's atomic SQL claim could hit WorktreeManager's
branch-collision self-heal, which force-removes and recreates the
winner's live worktree mid-run.

Move the claim before any resource creation and bail out immediately
when it's rejected. OverrideSlotService.RunNow also fast-rejects a
task already Running in the DB (defense in depth). RunCancellationRegistry
now refuses (and logs) a double registration instead of silently
overwriting the first CTS, so a losing dispatch's cleanup can no
longer unregister the winner's cancellation token.

src/ClaudeDo.Worker/Queue/OverrideSlotService.cs

@@ -33,9 +33,15 @@ public sealed class OverrideSlotService
     {
         using (var context = _dbFactory.CreateDbContext())
         {
-            var exists = await new TaskRepository(context).GetByIdAsync(taskId);
-            if (exists is null)
+            var task = await new TaskRepository(context).GetByIdAsync(taskId);
+            if (task is null)
                 throw new KeyNotFoundException($"Task '{taskId}' not found.");
+
+            // Fast-fail precheck only — not the race guard. A task sitting Queued still
+            // passes here and can race the queue picker's atomic claim; TaskRunner.RunAsync
+            // resolves that by claiming Running before creating any resources.
+            if (task.Status == Data.Models.TaskStatus.Running)
+                throw new InvalidOperationException("task is already running");
         }
 
         StartInSlot(taskId, ct => RunInSlotAsync(taskId, ct), "RunInSlotAsync failed for task {TaskId}");