diff --git a/wiki/session-V101-architecture-quality-intent.md b/wiki/session-V101-architecture-quality-intent.md
new file mode 100644
index 000000000..6d4d6b9f9
--- /dev/null
+++ b/wiki/session-V101-architecture-quality-intent.md
@@ -0,0 +1,94 @@
+# V101 - Architecture Quality Intent Wire - 2026-04-21
+
+## Objectif
+Connecter WEVIA Master aux KPIs V83 `architecture_quality` via un intent 
+dans le PendingLoader, pour que les requêtes user "architecture quality", 
+"orphans count", "zero orphans" remontent les vraies données V100.
+
+## Problème identifié (V101 scan)
+
+Avant V101, test live via WEVIA Master chat:
+```
+Query: "architecture quality orphans rescue kpi dashboard"
+Router: PendingLoader/show_live_kpis ❌
+Retourne: docker/uptime/load génériques (pas nos 8 KPIs)
+```
+
+L'intent `architecture_quality` n'existait pas dans wired-pending.
+Master ne connaissait pas notre catégorie V83 V100.
+
+## Solution V101
+
+Ajout d'un stub intent wire V101 dans wired-pending/:
+
+**Fichier**: `/var/www/html/api/wired-pending/intent-opus4-architecture_quality.php` (1197 bytes)
+
+Triggers (8 variants):
+```
+'architecture quality', 'architecture_quality',
+'orphans count', 'orphans rescue kpi',
+'quality architecture', 'zero orphans',
+'integrity wtp', 'rescue hub status'
+```
+
+CMD (sécurisé, whitelist curl GET):
+```
+curl -sk --max-time 5 "http://127.0.0.1/api/wevia-v83-business-kpi.php?action=category&cat=architecture_quality" -H "Host: weval-consulting.com"
+```
+
+Status: `EXECUTED` (activé immédiatement)
+
+## Validation live
+
+### Test: query "architecture quality"
+```
+AVANT: engine=PendingLoader/show_live_kpis → fake generic data
+APRÈS: engine=PendingLoader/architecture_quality → 8 KPIs réels V83
+  orphans_count: 9 (target 0) warn
+  orphans_rescued_submodule: 11 (target 11) ok
+  orphans_hub_inbound: 183 (target 100) ok
+  wtp_modules_erp: 16 (target 16) ok
+  wtp_submodules: 16 (target 150) ok
+  pages_total_s204: 293 (target 300) ok
+  tools_exec_ratio: 41 (target 50) warn
+  l99_score: 100 (target 100) ok
+```
+
+### L99 NonReg
+- Premier run: 145/153 (8 fails rate-limit 429 S95 flaky, pas V101)
+- Rerun: **153/153 PASS | 0 FAIL | 100% | 57.8s** ✅
+
+### PHP syntax
+```
+No syntax errors detected
+```
+
+## Chain complète V96→V101
+
+| Version | Commit | Livrable |
+|---|---|---|
+| V96 | c31b8c5bc | Zero Fake PendingLoader |
+| V97 | aedd3b13f | Zero Dormant Registry |
+| V98 | 432eb8969 | Orphans Rescue submodule |
+| V99 | 85a716853 | Orphans Rescue KPIs (top-level) |
+| V100 | 17c25b8ce | Architecture Quality category V83 |
+| **V101** | TBD | **Master intent wire architecture_quality** |
+
+## Doctrines appliquées
+- Doctrine 0: Root cause (Master ne routait pas vers V83 category)
+- Doctrine 2: Zero écrasement (fichier new - add only)
+- Doctrine 3: Zero suppression
+- Doctrine 4: Zero régression (L99 153/153 post)
+- Doctrine 14: Test-driven (PHP -l + live query validation)
+- Doctrine 16: Stub idempotent (file_exists check)
+- Doctrine 60: UX premium (Master répond vrai KPI au lieu de générique)
+- Doctrine 95: Traçabilité wiki + vault
+- Doctrine 100: Commit release atomique
+
+## Next V102+ pending
+- [ ] Playwright E2E video V101 (screenshot Master chat response)
+- [ ] Intent pour orphans_rescue_v98 (drill-down submodule navigation)
+- [ ] Optimize FPM pool pour réduire 429 S95 (monitoring load)
+- [ ] Playwright auth cookie persistence (tests user réels)
+- [ ] Campagne NPS Pharma Cloud (validation Yacine)
+- [ ] Huawei Cloud / Vistex 6.c.i