feat(dashboard): LLM stats answer who/what spends tokens — member scope chips and spend breakdowns

The stats section was designed single-user: four global charts, no
attribution. Request metadata rows now carry the session's source and
the frame's agent_id/depth (additive ensure_column), denormalized at log
time by the LoggingModel from the owner's pool — off the turn's hot
path, degrading to NULLs, never to a lost row.

The dashboard gains member scope chips filtering every chart, and a
breakdown row splitting the range's billed tokens by member, kind
(chat / sub-agents / cron / system agents / channels), agent, model and
provider — a sub-agent's spend is attributed to the sub-agent itself.
Rows predating the columns group under 'older data'.

web/i18n/fr.js

@@ -416,7 +416,6 @@ export default {
   'dashboard.stats.requests': 'Requêtes {per}',
   'dashboard.stats.tokens':   'Tokens {per}',
   'dashboard.stats.latency':  'Latence moyenne (ms)',
-  'dashboard.stats.models':   'Modèles',
   'dashboard.stats.per_min':  '/ min',
   'dashboard.stats.per_hour': '/ heure',
   'dashboard.stats.per_day':  '/ jour',
@@ -432,6 +431,26 @@ export default {
   'dashboard.stats.chart.non_cached': 'Non en cache',
   'dashboard.stats.chart.cache_hit':  'Cache hit : {pct}%',
 
+  'dashboard.stats.scope.all':      'Tous',
+  'dashboard.stats.sub.breakdown':  'Répartition des dépenses sur la période sélectionnée',
+  'dashboard.stats.by_user':        'Par membre',
+  'dashboard.stats.by_kind':        'Par type',
+  'dashboard.stats.by_agent':       'Par agent',
+  'dashboard.stats.by_model':       'Par modèle',
+  'dashboard.stats.by_provider':    'Par fournisseur',
+  'dashboard.stats.user_unknown':   'Inconnu',
+  'dashboard.stats.tip.requests':   'requêtes',
+
+  'dashboard.stats.kind.web':                 'Chat',
+  'dashboard.stats.kind.mobile':              'Mobile',
+  'dashboard.stats.kind.telegram':            'Telegram',
+  'dashboard.stats.kind.cron':                'Tâches planifiées',
+  'dashboard.stats.kind.sub_agent':           'Sous-agents',
+  'dashboard.stats.kind.event_triage':        'Triage des événements',
+  'dashboard.stats.kind.memory_lint':         'Lint mémoire',
+  'dashboard.stats.kind.conversation_review': 'Revue des conversations',
+  'dashboard.stats.kind.unknown':             'Données anciennes',
+
   'dashboard.hero.subtitle':  'Votre centre de commande IA — recherche, code, planification et orchestration. Tout en un seul endroit.',
 
   'dashboard.banner.no_models.title':  'Aucun modèle LLM configuré.',