feat(llm): 思考开关与思考强度拆成两个独立可配置参数

- thinking_type(''/disabled/enabled) 与 reasoning_effort(''/low/medium/high/xhigh/max)
  彻底解耦，可任意组合：支持「不发 thinking 字段、只靠强度激活思考」的接口
- agent.Config 拆为 ThinkingType + ReasoningEffort 两字段，NewProvider 直接透传
- db/schema.sql 加 thinking_type 列 + 一次性 DO 块迁移(仅列不存在时回填旧语义)
- 前端 LLM 配置页改为两个独立下拉，强度新增 xhigh 档

Claude-Session: https://claude.ai/code/session_01Bs2rrsVHUBTxRusm9espwB

agent/provider.go

@@ -41,9 +41,14 @@ type Config struct {
 	// ContextWindowK is the model's context window in K tokens (user-configured),
 	// used to size compaction thresholds. 0 = default; see CompactionWindow.
 	ContextWindowK int
-	// ReasoningEffort selects the thinking mode: "" = 默认(不发送思考参数);
-	// "off" = 关闭(thinking.type=disabled); "low"/"medium"/"high"/"max" = 开启并设强度.
-	// NewProvider derives the provider-specific request fields from it.
+	// ThinkingType 独立控制思考「开关」字段(thinking.type):
+	//   "" = 不发送(默认,兼容不支持该字段的模型); "disabled" = 显式关闭;
+	//   "enabled" = 开启. 与 ReasoningEffort 完全解耦——有些接口没有 thinking 字段、
+	//   只靠强度参数就能激活思考,故两者可各自单独设置.
+	ThinkingType string
+	// ReasoningEffort 独立控制思考「强度」字段:
+	//   "" = 不发送(默认); "low"/"medium"/"high"/"xhigh"/"max" = 对应强度.
+	//   OpenAI 映射为顶层 reasoning_effort;Anthropic 映射为 output_config.effort.
 	ReasoningEffort string
 }
 
@@ -184,17 +189,10 @@ func (c Config) NewProvider() (llm.Provider, error) {
 		Model:      c.Model,
 		HTTPClient: client,
 	}
-	// Derive the provider thinking params from the single UI selector:
-	//   "" → omit both (provider default); "off" → disable; else → enable + effort.
-	switch c.ReasoningEffort {
-	case "":
-		// leave both empty → omitted from the request
-	case "off":
-		lc.ThinkingType = "disabled"
-	default:
-		lc.ThinkingType = "enabled"
-		lc.ReasoningEffort = c.ReasoningEffort
-	}
+	// 思考开关与强度两个字段各自透传(空 = 该字段不发送)。二者解耦:
+	// 可只发 thinking.type、只发 effort、都发、或都不发。
+	lc.ThinkingType = c.ThinkingType
+	lc.ReasoningEffort = c.ReasoningEffort
 	if c.RatePerSecond > 0 || c.RatePerMinute > 0 {
 		lc.RateLimit = &llm.RateLimit{PerSecond: c.RatePerSecond, PerMinute: c.RatePerMinute}
 	}