feat(plugin, api): introduce host model callback support with Go example and API handlers

- Added an example plugin `host-model-callback` in Go to summarize host model callbacks.
- Implemented `cliproxy_plugin_init`, `cliproxyPluginCall`, and other plugin functions for callback handling.
- Introduced API handlers for `ModelExecution` and `ModelExecutionStream` with support for both streaming and non-streaming requests.
- Included unit tests (`model_execution_test.go`) to validate execution logic and streaming responses.

internal/runtime/executor/claude_executor.go

@@ -174,6 +174,7 @@ func (e *ClaudeExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, r
 	reporter := helps.NewExecutorUsageReporter(ctx, e, baseModel, auth)
 	defer reporter.TrackFailure(ctx, &err)
 	from := opts.SourceFormat
+	responseFormat := cliproxyexecutor.ResponseFormatOrSource(opts)
 	to := sdktranslator.FromString("claude")
 	// Use streaming translation to preserve function calling, except for claude.
 	stream := from != to
@@ -332,7 +333,7 @@ func (e *ClaudeExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, r
 	out := sdktranslator.TranslateNonStream(
 		ctx,
 		to,
-		from,
+		responseFormat,
 		req.Model,
 		opts.OriginalRequest,
 		bodyForTranslation,
@@ -357,6 +358,7 @@ func (e *ClaudeExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 	reporter := helps.NewExecutorUsageReporter(ctx, e, baseModel, auth)
 	defer reporter.TrackFailure(ctx, &err)
 	from := opts.SourceFormat
+	responseFormat := cliproxyexecutor.ResponseFormatOrSource(opts)
 	to := sdktranslator.FromString("claude")
 	originalPayloadSource := req.Payload
 	if len(opts.OriginalRequest) > 0 {
@@ -488,8 +490,8 @@ func (e *ClaudeExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 			}
 		}()
 
-		// If from == to (Claude → Claude), directly forward the SSE stream without translation
-		if from == to {
+		// If the response target is Claude, directly forward the SSE stream without translation.
+		if responseFormat == to {
 			scanner := bufio.NewScanner(decodedBody)
 			scanner.Buffer(nil, 52_428_800) // 50MB
 			for scanner.Scan() {
@@ -534,7 +536,7 @@ func (e *ClaudeExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.A
 			chunks := sdktranslator.TranslateStream(
 				ctx,
 				to,
-				from,
+				responseFormat,
 				req.Model,
 				opts.OriginalRequest,
 				bodyForTranslation,
@@ -628,6 +630,7 @@ func (e *ClaudeExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Aut
 	}
 
 	from := opts.SourceFormat
+	responseFormat := cliproxyexecutor.ResponseFormatOrSource(opts)
 	to := sdktranslator.FromString("claude")
 	// Use streaming translation to preserve function calling, except for claude.
 	stream := from != to
@@ -725,7 +728,7 @@ func (e *ClaudeExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Aut
 	}
 	helps.AppendAPIResponseChunk(ctx, e.cfg, data)
 	count := gjson.GetBytes(data, "input_tokens").Int()
-	out := sdktranslator.TranslateTokenCount(ctx, to, from, count, data)
+	out := sdktranslator.TranslateTokenCount(ctx, to, responseFormat, count, data)
 	return cliproxyexecutor.Response{Payload: out, Headers: resp.Header.Clone()}, nil
 }