feat(model): 双协议 LLM Provider（AnthropicChNativeAdapter /v1/messages + tool_use.input→Args 直接映射 + 流式重组）+ Provider 感知工厂/路由 + LLM_PROVIDER/anthropic 配置（OpenAI 兼容保留）

internal/runtime/model/model.go

@@ -87,6 +87,18 @@ type ToolSpec struct {
 	Parameters any
 }
 
+// ToolCall is a single assistant-originated tool invocation parsed from a
+// provider's tool_use content block. Args is the structured argument object
+// (never a serialized JSON string), so a tool loop can bind it directly without
+// re-parsing.
+type ToolCall struct {
+	// Name is the tool being invoked.
+	Name string
+	// Args are the tool arguments, decoded from the provider's structured
+	// tool_use.input object. It is never a JSON-encoded string.
+	Args map[string]any
+}
+
 // Request is a completion request, normalized across providers.
 type Request struct {
 	// Model overrides the adapter's configured default model. Empty means the
@@ -101,6 +113,11 @@ type Request struct {
 	Purpose string
 	// Tools is the interface-level placeholder for M2 function calling.
 	Tools []ToolSpec
+	// ToolChoice optionally directs how the provider selects a tool. It is an
+	// opaque provider-agnostic directive (e.g. {"type":"auto"} or {"type":"tool",
+	// "name":"x"}); empty means the provider default. Passed through verbatim to
+	// protocols that support it (Anthropic's tool_choice).
+	ToolChoice any
 }
 
 // Response is a normalized completion result.
@@ -110,6 +127,9 @@ type Response struct {
 	Content string
 	// Message is the normalized assistant message.
 	Message Message
+	// ToolCalls carries any tool_use blocks the assistant emitted. Empty when the
+	// response is pure text.
+	ToolCalls []ToolCall
 	// PromptTokens, CompletionTokens and TotalTokens report provider usage when
 	// available; zero when the provider does not return it.
 	PromptTokens     int