feat(Phase 3): 學習閉環補完 — Root cause 3 + 診斷 feedback + 知識遺忘 + Fine-tune 管線

- approval_execution.py: _run_post_execution_verify() 補接 record_verification_result()
  Root cause 3 終結：環境驗證結果（success/degraded/failed/timeout）不再孤立
- learning_service.py: 新增 record_verification_result() — 驗證結果 → Redis + Playbook EWMA
- learning_service.py: 新增 record_diagnosis_outcome() — 誤診負向訊號回寫（L3×D4）
- jobs/knowledge_decay_job.py: 新建 30d 知識遺忘 Job（未引用 draft/review → archived）
- services/finetune_exporter.py: 新建每週 JSONL 匯出（EvidenceSnapshot × AgentSession）
- main.py: 掛載 knowledge_decay_loop（24h）+ finetune_export_loop（7d）
- MASTER §8: Phase 3 核心改造項全部落地記錄

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

apps/api/src/services/approval_execution.py

@@ -568,6 +568,25 @@ class ApprovalExecutionService:
                 action=action_taken,
             )
 
+            # ADR-083 Phase 3 Root cause 3: 驗證結果接線到學習服務
+            # 環境驗證（Pod Running / 指標恢復）是比執行 exit code 更精確的學習訊號
+            # 2026-04-15 ogt + Claude Sonnet 4.6（亞太）
+            try:
+                from src.services.learning_service import get_learning_service
+                _matched_pb_id = getattr(approval, "matched_playbook_id", None)
+                await get_learning_service().record_verification_result(
+                    incident_id=approval.incident_id,
+                    action_taken=action_taken,
+                    verification_result=verification_result,
+                    matched_playbook_id=_matched_pb_id,
+                )
+            except Exception as _lerr:
+                logger.warning(
+                    "post_verify_learning_failed",
+                    approval_id=str(approval.id),
+                    error=str(_lerr),
+                )
+
         except Exception as _e:
             # 驗證失敗不影響執行結果
             logger.warning(