diff --git a/docs/LOGBOOK.md b/docs/LOGBOOK.md
index 3a738b70..b7090453 100644
--- a/docs/LOGBOOK.md
+++ b/docs/LOGBOOK.md
@@ -6,6 +6,17 @@
 
 ---
 
+## 2026-05-05 | drift-scanner CronJob 納入 ArgoCD baseline
+
+**背景**：重開機恢復後，K8s Deployments 與三個新納入的 CronJob 已跟到最新 image，但 `drift-scanner` 仍是手動套用的舊固定 SHA，會造成「服務健康、排程吃舊版」的冷啟動盲區。
+
+**本次修補**：
+- 將 `drift-scanner` manifest 移入 `k8s/awoooi-prod/12-cronjob-drift-scanner.yaml`，由 `k8s/awoooi-prod/kustomization.yaml` 納入 ArgoCD 管理。
+- `drift-scanner` image 改用 `192.168.0.110:5000/library/api:IMAGE_TAG_PLACEHOLDER`，讓 CD 的 kustomize image 注入同時覆蓋 drift 排程。
+
+**驗證**：
+- `kubectl kustomize k8s/awoooi-prod` 通過，build output 中 `drift-scanner` image 會被解析為目前 kustomization 的 `awoooi/api:c4854bb3...`。
+
 ## 2026-05-05 | 重開機後排程與 startup baseline 修復
 
 **背景**：四台主機非預期重開機後，統帥要求確認所有服務、網站、工具、資料庫與排程都能正常恢復，不能只看容器 `healthy`。
diff --git a/k8s/drift-cronjob.yaml b/k8s/awoooi-prod/12-cronjob-drift-scanner.yaml
similarity index 88%
rename from k8s/drift-cronjob.yaml
rename to k8s/awoooi-prod/12-cronjob-drift-scanner.yaml
index 7bf44a0f..e8191f13 100644
--- a/k8s/drift-cronjob.yaml
+++ b/k8s/awoooi-prod/12-cronjob-drift-scanner.yaml
@@ -6,7 +6,7 @@
 # 關聯設計: docs/superpowers/specs/2026-04-04-nemotron-active-defense-design.md 方向三
 # 關聯 ADR: 待起草 ADR-057
 #
-# 部署: kubectl apply -f k8s/drift-cronjob.yaml -n awoooi-prod
+# 部署: 由 ArgoCD 套用 k8s/awoooi-prod/kustomization.yaml
 # 手動觸發: kubectl create job --from=cronjob/drift-scanner drift-scan-manual -n awoooi-prod
 # 查看 log: kubectl logs -l job-name=drift-scanner -n awoooi-prod
 
@@ -48,8 +48,11 @@ spec:
           containers:
             - name: drift-scanner
               # 使用 awoooi-api 鏡像（含 kubectl + Python 環境）
-              # 2026-04-09 Claude Sonnet 4.6: 改用內網 registry + 固定 SHA tag (禁止 latest)
-              image: 192.168.0.110:5000/awoooi/api:21567a7a6dbee7db2c0f59c265f80713ff5e6fe4
+              # 2026-05-05 Codex: keep the API image placeholder so CD
+              # injects the same immutable tag used by API/worker. Leaving
+              # this CronJob on a fixed old tag made the schedule drift after
+              # reboot even when Deployments were healthy.
+              image: 192.168.0.110:5000/library/api:IMAGE_TAG_PLACEHOLDER
               imagePullPolicy: Always
               command:
                 - python
diff --git a/k8s/awoooi-prod/kustomization.yaml b/k8s/awoooi-prod/kustomization.yaml
index 9ee862be..70d259c4 100644
--- a/k8s/awoooi-prod/kustomization.yaml
+++ b/k8s/awoooi-prod/kustomization.yaml
@@ -31,6 +31,7 @@ resources:
 - 13-cronjob-k3s-report.yaml
 - 14-cronjob-weekly-report.yaml
 - 15-cronjob-km-vectorize.yaml
+- 12-cronjob-drift-scanner.yaml
 
 # 映像配置 (Tag 由 CI 動態注入)
 # Harbor 金庫: 110 主機 (192.168.0.110:5000)