refactor(hermes-auto-improve): scheduling via n8n — supprimer timer systemd

Le scheduling est géré par n8n (workflow "Hermes Doc Worker — Funk-lab") :
- Schedule 30min désactivé dans n8n
- Schedule 22h → POST /trigger/all (tous les docs, timeout 2h)
- Attendre run 5min → 120min dans le workflow n8n

Ansible : retirer les templates daily service/timer (plus nécessaires)
auto-improve.py : flag --all conservé (utilisé par le trigger /trigger/all)
trigger-server.py : timeout 7200s conservé pour les runs --all

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

ansible/roles/hermes_auto_improve/handlers/main.yml

@@ -5,10 +5,6 @@
     state: restarted
     daemon_reload: true
 
-- name: Reload systemd daemon
-  ansible.builtin.systemd:
-    daemon_reload: true
-
 - name: Reload nftables
   ansible.builtin.service:
     name: nftables

ansible/roles/hermes_auto_improve/tasks/main.yml

@@ -47,24 +47,3 @@
     state: started
     daemon_reload: true
 
-- name: Deploy daily service (oneshot — analyse complète)
-  ansible.builtin.template:
-    src: hermes-auto-improve-daily.service.j2
-    dest: /etc/systemd/system/hermes-auto-improve-daily.service
-    mode: '0644'
-  notify: Reload systemd daemon
-
-- name: Deploy daily timer (22h00)
-  ansible.builtin.template:
-    src: hermes-auto-improve-daily.timer.j2
-    dest: /etc/systemd/system/hermes-auto-improve-daily.timer
-    mode: '0644'
-  notify: Reload systemd daemon
-
-- name: Enable and start daily timer
-  ansible.builtin.systemd:
-    name: hermes-auto-improve-daily.timer
-    enabled: true
-    state: started
-    daemon_reload: true
-