Add metadata to create message (#3832)

.changeset/hot-lies-flash.md

@@ -0,0 +1,5 @@
+---
+"roo-cline": patch
+---
+
+Add metadata to create message

src/api/index.ts

@@ -29,8 +29,17 @@ export interface SingleCompletionHandler {
 	completePrompt(prompt: string): Promise<string>
 }
 
+export interface ApiHandlerCreateMessageMetadata {
+	mode?: string
+	taskId: string
+}
+
 export interface ApiHandler {
-	createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream
+	createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream
 
 	getModel(): { id: string; info: ModelInfo }
 

src/api/providers/anthropic-vertex.ts

@@ -11,7 +11,7 @@ import { getModelParams } from "../transform/model-params"
 
 import { ANTHROPIC_DEFAULT_MAX_TOKENS } from "./constants"
 import { BaseProvider } from "./base-provider"
-import type { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 // https://docs.anthropic.com/en/api/claude-on-vertex-ai
 export class AnthropicVertexHandler extends BaseProvider implements SingleCompletionHandler {
@@ -50,7 +50,11 @@ export class AnthropicVertexHandler extends BaseProvider implements SingleComple
 		}
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		let {
 			id,
 			info: { supportsPromptCache },

src/api/providers/anthropic.ts

@@ -15,7 +15,7 @@ import { getModelParams } from "../transform/model-params"
 
 import { ANTHROPIC_DEFAULT_MAX_TOKENS } from "./constants"
 import { BaseProvider } from "./base-provider"
-import type { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 export class AnthropicHandler extends BaseProvider implements SingleCompletionHandler {
 	private options: ApiHandlerOptions
@@ -34,7 +34,11 @@ export class AnthropicHandler extends BaseProvider implements SingleCompletionHa
 		})
 	}
 
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		let stream: AnthropicStream<Anthropic.Messages.RawMessageStreamEvent>
 		const cacheControl: CacheControlEphemeral = { type: "ephemeral" }
 		let { id: modelId, betas = [], maxTokens, temperature, reasoning: thinking } = this.getModel()

src/api/providers/base-openai-compatible-provider.ts

@@ -5,7 +5,7 @@ import { ApiHandlerOptions, ModelInfo } from "../../shared/api"
 import { ApiStream } from "../transform/stream"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 
-import { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { DEFAULT_HEADERS } from "./constants"
 import { BaseProvider } from "./base-provider"
 
@@ -60,7 +60,11 @@ export abstract class BaseOpenAiCompatibleProvider<ModelName extends string>
 		})
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const {
 			id: model,
 			info: { maxTokens: max_tokens },

src/api/providers/base-provider.ts

@@ -2,7 +2,7 @@ import { Anthropic } from "@anthropic-ai/sdk"
 
 import { ModelInfo } from "../../shared/api"
 
-import { ApiHandler } from "../index"
+import type { ApiHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { ApiStream } from "../transform/stream"
 import { countTokens } from "../../utils/countTokens"
 
@@ -10,7 +10,11 @@ import { countTokens } from "../../utils/countTokens"
  * Base class for API providers that implements common functionality.
  */
 export abstract class BaseProvider implements ApiHandler {
-	abstract createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream
+	abstract createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream
 	abstract getModel(): { id: string; info: ModelInfo }
 
 	/**

src/api/providers/bedrock.ts

@@ -9,7 +9,6 @@ import {
 } from "@aws-sdk/client-bedrock-runtime"
 import { fromIni } from "@aws-sdk/credential-providers"
 import { Anthropic } from "@anthropic-ai/sdk"
-import { SingleCompletionHandler } from "../"
 import {
 	BedrockModelId,
 	ModelInfo as SharedModelInfo,
@@ -26,6 +25,7 @@ import { MultiPointStrategy } from "../transform/cache-strategy/multi-point-stra
 import { ModelInfo as CacheModelInfo } from "../transform/cache-strategy/types"
 import { AMAZON_BEDROCK_REGION_INFO } from "../../shared/aws_regions"
 import { convertToBedrockConverseMessages as sharedConverter } from "../transform/bedrock-converse-format"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 const BEDROCK_DEFAULT_TEMPERATURE = 0.3
 const BEDROCK_MAX_TOKENS = 4096
@@ -189,7 +189,11 @@ export class AwsBedrockHandler extends BaseProvider implements SingleCompletionH
 		this.client = new BedrockRuntimeClient(clientConfig)
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		let modelConfig = this.getModel()
 		// Handle cross-region inference
 		const usePromptCache = Boolean(this.options.awsUsePromptCache && this.supportsAwsPromptCache(modelConfig))
@@ -769,7 +773,7 @@ export class AwsBedrockHandler extends BaseProvider implements SingleCompletionH
 	> = {
 		ACCESS_DENIED: {
 			patterns: ["access", "denied", "permission"],
-			messageTemplate: `You don't have access to the model specified. 
+			messageTemplate: `You don't have access to the model specified.
 
 Please verify:
 1. Try cross-region inference if you're using a foundation model

src/api/providers/fake-ai.ts

@@ -1,7 +1,7 @@
 import { Anthropic } from "@anthropic-ai/sdk"
-import { ApiHandler, SingleCompletionHandler } from ".."
 import { ApiHandlerOptions, ModelInfo } from "../../shared/api"
 import { ApiStream } from "../transform/stream"
+import type { ApiHandler, SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 interface FakeAI {
 	/**
@@ -18,7 +18,11 @@ interface FakeAI {
 	 */
 	removeFromCache?: () => void
 
-	createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream
+	createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream
 	getModel(): { id: string; info: ModelInfo }
 	countTokens(content: Array<Anthropic.Messages.ContentBlockParam>): Promise<number>
 	completePrompt(prompt: string): Promise<string>
@@ -52,8 +56,12 @@ export class FakeAIHandler implements ApiHandler, SingleCompletionHandler {
 		this.ai = cachedFakeAi
 	}
 
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
-		yield* this.ai.createMessage(systemPrompt, messages)
+	async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
+		yield* this.ai.createMessage(systemPrompt, messages, metadata)
 	}
 
 	getModel(): { id: string; info: ModelInfo } {

src/api/providers/gemini.ts

@@ -10,7 +10,7 @@ import type { JWTInput } from "google-auth-library"
 import { ApiHandlerOptions, ModelInfo, GeminiModelId, geminiDefaultModelId, geminiModels } from "../../shared/api"
 import { safeJsonParse } from "../../shared/safeJsonParse"
 
-import { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { convertAnthropicContentToGemini, convertAnthropicMessageToGemini } from "../transform/gemini-format"
 import type { ApiStream } from "../transform/stream"
 import { BaseProvider } from "./base-provider"
@@ -54,7 +54,11 @@ export class GeminiHandler extends BaseProvider implements SingleCompletionHandl
 					: new GoogleGenAI({ apiKey })
 	}
 
-	async *createMessage(systemInstruction: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(
+		systemInstruction: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const { id: model, thinkingConfig, maxOutputTokens, info } = this.getModel()
 
 		const contents = messages.map(convertAnthropicMessageToGemini)

src/api/providers/glama.ts

@@ -9,7 +9,7 @@ import { ApiStream } from "../transform/stream"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { addCacheBreakpoints } from "../transform/caching/anthropic"
 
-import { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { RouterProvider } from "./router-provider"
 
 const GLAMA_DEFAULT_TEMPERATURE = 0
@@ -33,7 +33,11 @@ export class GlamaHandler extends RouterProvider implements SingleCompletionHand
 		})
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const { id: modelId, info } = await this.fetchModel()
 
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/api/providers/human-relay.ts

@@ -4,7 +4,7 @@ import * as vscode from "vscode"
 import { ModelInfo } from "../../shared/api"
 import { getCommand } from "../../utils/commands"
 import { ApiStream } from "../transform/stream"
-import { ApiHandler, SingleCompletionHandler } from "../index"
+import type { ApiHandler, SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 /**
  * Human Relay API processor
  * This processor does not directly call the API, but interacts with the model through human operations copy and paste.
@@ -18,8 +18,13 @@ export class HumanRelayHandler implements ApiHandler, SingleCompletionHandler {
 	 * Create a message processing flow, display a dialog box to request human assistance
 	 * @param systemPrompt System prompt words
 	 * @param messages Message list
+	 * @param metadata Optional metadata
 	 */
-	async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		// Get the most recent user message
 		const latestMessage = messages[messages.length - 1]
 

src/api/providers/litellm.ts

@@ -4,7 +4,7 @@ import { Anthropic } from "@anthropic-ai/sdk" // Keep for type usage only
 import { ApiHandlerOptions, litellmDefaultModelId, litellmDefaultModelInfo } from "../../shared/api"
 import { ApiStream, ApiStreamUsageChunk } from "../transform/stream"
 import { convertToOpenAiMessages } from "../transform/openai-format"
-import { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { RouterProvider } from "./router-provider"
 
 /**
@@ -26,7 +26,11 @@ export class LiteLLMHandler extends RouterProvider implements SingleCompletionHa
 		})
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const { id: modelId, info } = await this.fetchModel()
 
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/api/providers/lmstudio.ts

@@ -2,12 +2,12 @@ import { Anthropic } from "@anthropic-ai/sdk"
 import OpenAI from "openai"
 import axios from "axios"
 
-import { SingleCompletionHandler } from "../"
 import { ApiHandlerOptions, ModelInfo, openAiModelInfoSaneDefaults } from "../../shared/api"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
 import { BaseProvider } from "./base-provider"
 import { XmlMatcher } from "../../utils/xml-matcher"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 const LMSTUDIO_DEFAULT_TEMPERATURE = 0
 
@@ -24,7 +24,11 @@ export class LmStudioHandler extends BaseProvider implements SingleCompletionHan
 		})
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
 			{ role: "system", content: systemPrompt },
 			...convertToOpenAiMessages(messages),

src/api/providers/mistral.ts

@@ -1,10 +1,10 @@
 import { Anthropic } from "@anthropic-ai/sdk"
 import { Mistral } from "@mistralai/mistralai"
-import { SingleCompletionHandler } from "../"
 import { ApiHandlerOptions, mistralDefaultModelId, MistralModelId, mistralModels, ModelInfo } from "../../shared/api"
 import { convertToMistralMessages } from "../transform/mistral-format"
 import { ApiStream } from "../transform/stream"
 import { BaseProvider } from "./base-provider"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 const MISTRAL_DEFAULT_TEMPERATURE = 0
 
@@ -41,7 +41,13 @@ export class MistralHandler extends BaseProvider implements SingleCompletionHand
 		return "https://api.mistral.ai"
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
+		const { id: model } = this.getModel()
+
 		const response = await this.client.chat.stream({
 			model: this.options.apiModelId || mistralDefaultModelId,
 			messages: [{ role: "system", content: systemPrompt }, ...convertToMistralMessages(messages)],

src/api/providers/ollama.ts

@@ -2,7 +2,7 @@ import { Anthropic } from "@anthropic-ai/sdk"
 import OpenAI from "openai"
 import axios from "axios"
 
-import { SingleCompletionHandler } from "../"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { ApiHandlerOptions, ModelInfo, openAiModelInfoSaneDefaults } from "../../shared/api"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { convertToR1Format } from "../transform/r1-format"
@@ -27,7 +27,11 @@ export class OllamaHandler extends BaseProvider implements SingleCompletionHandl
 		})
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const modelId = this.getModel().id
 		const useR1Format = modelId.toLowerCase().includes("deepseek-r1")
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/api/providers/openai-native.ts

@@ -15,7 +15,7 @@ import { convertToOpenAiMessages } from "../transform/openai-format"
 import { ApiStream } from "../transform/stream"
 import { getModelParams } from "../transform/model-params"
 
-import type { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { BaseProvider } from "./base-provider"
 
 const OPENAI_NATIVE_DEFAULT_TEMPERATURE = 0
@@ -33,7 +33,11 @@ export class OpenAiNativeHandler extends BaseProvider implements SingleCompletio
 		this.client = new OpenAI({ baseURL: this.options.openAiNativeBaseUrl, apiKey })
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const model = this.getModel()
 		let id: "o3-mini" | "o3" | "o4-mini" | undefined
 

src/api/providers/openai.ts

@@ -18,7 +18,7 @@ import { ApiStream, ApiStreamUsageChunk } from "../transform/stream"
 import { getModelParams } from "../transform/model-params"
 
 import { DEFAULT_HEADERS, DEEP_SEEK_DEFAULT_TEMPERATURE } from "./constants"
-import type { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { BaseProvider } from "./base-provider"
 
 export const AZURE_AI_INFERENCE_PATH = "/models/chat/completions"
@@ -71,7 +71,11 @@ export class OpenAiHandler extends BaseProvider implements SingleCompletionHandl
 		}
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const { info: modelInfo, reasoning } = this.getModel()
 		const modelUrl = this.options.openAiBaseUrl ?? ""
 		const modelId = this.options.openAiModelId ?? ""

src/api/providers/requesty.ts

@@ -9,7 +9,7 @@ import {
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { calculateApiCostOpenAI } from "../../utils/cost"
 import { ApiStream, ApiStreamUsageChunk } from "../transform/stream"
-import { SingleCompletionHandler } from "../"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../"
 import { BaseProvider } from "./base-provider"
 import { DEFAULT_HEADERS } from "./constants"
 import { getModels } from "./fetchers/modelCache"
@@ -25,7 +25,14 @@ interface RequestyUsage extends OpenAI.CompletionUsage {
 	total_cost?: number
 }
 
-type RequestyChatCompletionParams = OpenAI.Chat.ChatCompletionCreateParams & {}
+type RequestyChatCompletionParams = OpenAI.Chat.ChatCompletionCreateParams & {
+	requesty?: {
+		trace_id?: string
+		extra?: {
+			mode?: string
+		}
+	}
+}
 
 export class RequestyHandler extends BaseProvider implements SingleCompletionHandler {
 	protected options: ApiHandlerOptions
@@ -75,7 +82,11 @@ export class RequestyHandler extends BaseProvider implements SingleCompletionHan
 		}
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const model = await this.fetchModel()
 
 		let openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [
@@ -97,6 +108,12 @@ export class RequestyHandler extends BaseProvider implements SingleCompletionHan
 			temperature: temperature,
 			stream: true,
 			stream_options: { include_usage: true },
+			requesty: {
+				trace_id: metadata?.taskId,
+				extra: {
+					mode: metadata?.mode,
+				},
+			},
 		}
 
 		const stream = await this.client.chat.completions.create(completionParams)

src/api/providers/unbound.ts

@@ -7,7 +7,7 @@ import { ApiStream, ApiStreamUsageChunk } from "../transform/stream"
 import { convertToOpenAiMessages } from "../transform/openai-format"
 import { addCacheBreakpoints } from "../transform/caching/anthropic"
 
-import { SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 import { RouterProvider } from "./router-provider"
 
 const DEFAULT_HEADERS = {
@@ -32,7 +32,11 @@ export class UnboundHandler extends RouterProvider implements SingleCompletionHa
 		})
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const { id: modelId, info } = await this.fetchModel()
 
 		const openAiMessages: OpenAI.Chat.ChatCompletionMessageParam[] = [

src/api/providers/vscode-lm.ts

@@ -1,12 +1,12 @@
 import { Anthropic } from "@anthropic-ai/sdk"
 import * as vscode from "vscode"
 
-import { SingleCompletionHandler } from "../"
 import { ApiStream } from "../transform/stream"
 import { convertToVsCodeLmMessages } from "../transform/vscode-lm-format"
 import { SELECTOR_SEPARATOR, stringifyVsCodeLmModelSelector } from "../../shared/vsCodeSelectorUtils"
 import { ApiHandlerOptions, ModelInfo, openAiModelInfoSaneDefaults } from "../../shared/api"
 import { BaseProvider } from "./base-provider"
+import type { SingleCompletionHandler, ApiHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 /**
  * Handles interaction with VS Code's Language Model API for chat-based operations.
@@ -148,6 +148,7 @@ export class VsCodeLmHandler extends BaseProvider implements SingleCompletionHan
 	 *
 	 * @param systemPrompt - The system prompt to initialize the conversation context
 	 * @param messages - An array of message parameters following the Anthropic message format
+	 * @param metadata - Optional metadata for the message
 	 *
 	 * @yields {ApiStream} An async generator that yields either text chunks or tool calls from the model response
 	 *
@@ -329,7 +330,11 @@ export class VsCodeLmHandler extends BaseProvider implements SingleCompletionHan
 		return content
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		// Ensure clean state before starting a new request
 		this.ensureCleanState()
 		const client: vscode.LanguageModelChat = await this.getClient()

src/api/providers/xai.ts

@@ -9,7 +9,7 @@ import { getModelParams } from "../transform/model-params"
 
 import { DEFAULT_HEADERS } from "./constants"
 import { BaseProvider } from "./base-provider"
-import { type SingleCompletionHandler } from "../index"
+import type { SingleCompletionHandler, ApiHandlerCreateMessageMetadata } from "../index"
 
 const XAI_DEFAULT_TEMPERATURE = 0
 
@@ -38,7 +38,11 @@ export class XAIHandler extends BaseProvider implements SingleCompletionHandler
 		return { id, info, ...params }
 	}
 
-	override async *createMessage(systemPrompt: string, messages: Anthropic.Messages.MessageParam[]): ApiStream {
+	override async *createMessage(
+		systemPrompt: string,
+		messages: Anthropic.Messages.MessageParam[],
+		metadata?: ApiHandlerCreateMessageMetadata,
+	): ApiStream {
 		const { id: modelId, info: modelInfo, reasoning } = this.getModel()
 
 		// Use the OpenAI-compatible API.

src/core/task/Task.ts

@@ -12,7 +12,7 @@ import { serializeError } from "serialize-error"
 import { TokenUsage, ToolUsage, ToolName, ContextCondense } from "../../schemas"
 
 // api
-import { ApiHandler, buildApiHandler } from "../../api"
+import { ApiHandler, ApiHandlerCreateMessageMetadata, buildApiHandler } from "../../api"
 import { ApiStream } from "../../api/transform/stream"
 
 // shared
@@ -1499,6 +1499,7 @@ export class Task extends EventEmitter<ClineEvents> {
 			alwaysApproveResubmit,
 			requestDelaySeconds,
 			experiments,
+			mode,
 			autoCondenseContextPercent = 100,
 		} = state ?? {}
 
@@ -1614,7 +1615,12 @@ export class Task extends EventEmitter<ClineEvents> {
 			}
 		}
 
-		const stream = this.api.createMessage(systemPrompt, cleanConversationHistory)
+		const metadata: ApiHandlerCreateMessageMetadata = {
+			mode: mode,
+			taskId: this.taskId,
+		}
+
+		const stream = this.api.createMessage(systemPrompt, cleanConversationHistory, metadata)
 		const iterator = stream[Symbol.asyncIterator]()
 
 		try {