feat: add request dedup, response cache, and capability routing metrics to API server

- Add response cache middleware (backed by aster-rust implementation)
- Add request deduplication middleware to prevent duplicate in-flight requests
- Add capability routing metrics middleware for model/provider fallback tracking
- Add idempotency stats (atomic counters) to IdempotencyStore
- Expose all four stores in ServerState and surface stats in ServerStatus
- Add ResponseCacheSettings to ServerConfig with defaults (600s TTL, 200 max entries)
- Add API server observability infrastructure (IdempotencyGuard, RequestDedupGuard, ResponseCacheGuard)
- Add inline capability detection for vision/tools/context when routing requests
- Register workspace_ensure_ready and workspace_ensure_default_ready commands in runner
- Update docs with response_cache configuration example

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

docs/content/02.user-guide/4.configuration-example.md

@@ -72,6 +72,25 @@ routing:
 
 适用：有脚本联动、自动化流程需求的用户。
 
+## 示例 4：API 缓存策略（高级）
+
+目标：精确控制哪些响应状态码参与短时缓存（非流式）。
+
+```yaml
+server:
+  host: "127.0.0.1"
+  port: 8999
+  api_key: "your-api-key"
+  response_cache:
+    enabled: true
+    ttl_secs: 600
+    max_entries: 200
+    max_body_bytes: 1048576
+    cacheable_status_codes: [200]  # 默认仅缓存 200；可按需扩展 [200, 201]
+```
+
+适用：对缓存命中与语义一致性有要求的自动化/API 调用场景。
+
 ## 调整顺序建议
 
 1. 先确认导航与主题