merge: 合并 test 分支到 test-dev，解决冲突

解决的冲突文件：
- wire_gen.go: 合并 ConcurrencyService/CRSSyncService 参数和 userAttributeHandler
- gateway_handler.go: 合并 pkg/errors 和 antigravity 导入
- gateway_service.go: 合并 validateUpstreamBaseURL 和 GetAvailableModels
- config.example.yaml: 合并 billing/turnstile 配置和额外 gateway 选项

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

deploy/config.example.yaml

@@ -90,6 +90,14 @@ gateway:
   client_idle_ttl_seconds: 900
   # 并发槽位过期时间（分钟）
   concurrency_slot_ttl_minutes: 15
+  # Log upstream error response body summary (safe/truncated; does not log request content)
+  log_upstream_error_body: false
+  # Max bytes to log from upstream error body
+  log_upstream_error_body_max_bytes: 2048
+  # Auto inject anthropic-beta for API-key accounts when needed (default off)
+  inject_beta_for_apikey: false
+  # Allow failover on selected 400 errors (default off)
+  failover_on_400: false
 
 # =============================================================================
 # Database Configuration (PostgreSQL)
@@ -202,3 +210,19 @@ gemini:
     client_secret: "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"
     # Optional scopes (space-separated). Leave empty to auto-select based on oauth_type.
     scopes: ""
+  quota:
+    # Optional: local quota simulation for Gemini Code Assist (local billing).
+    # These values are used for UI progress + precheck scheduling, not official Google quotas.
+    tiers:
+      LEGACY:
+        pro_rpd: 50
+        flash_rpd: 1500
+        cooldown_minutes: 30
+      PRO:
+        pro_rpd: 1500
+        flash_rpd: 4000
+        cooldown_minutes: 5
+      ULTRA:
+        pro_rpd: 2000
+        flash_rpd: 0
+        cooldown_minutes: 5