feat(api): pass abortSignal to streaming API calls for all providers

- Add AbortController signal support across all OpenAI-compatible providers
  (openai, openrouter, lm-studio, deepseek, fireworks, vercel-ai-gateway,
   sambanova, unbound, zai, requesty, opencode-go, qwen-code, mimo, lite-llm)
- Add abort signal support to base-openai-compatible-provider
- Pass abort signal through Task class for cancellation propagation
- Add comprehensive abort signal tests for all providers
- Remove ripgrep diagnostic feature (Show Ripgrep Diagnostic command)
  - Clean up ripgrep service files and related tests
- Update list-files to check cwd before invoking rg
- Remove unused dependencies from package.json

Author: easonliang28

knip.json

@@ -21,7 +21,6 @@
 				"@types/node-cache",
 				"@types/vscode",
 				"@vscode/codicons",
-				"@vscode/ripgrep",
 				"esbuild-wasm",
 				"sambanova-ai-provider",
 				"tree-sitter-wasms",

packages/types/src/vscode.ts

@@ -46,8 +46,6 @@ export const commandIds = [
 	"acceptInput",
 	"focusPanel",
 	"toggleAutoApprove",
-
-	"showRipgrepDiagnostic",
 ] as const
 
 export type CommandId = (typeof commandIds)[number]

src/activate/__tests__/registerCommands.spec.ts

@@ -81,10 +81,6 @@ vi.mock("../../i18n", () => ({
 	t: (key: string) => key,
 }))
 
-vi.mock("../../services/ripgrep/diagnostic", () => ({
-	registerRipgrepDiagnosticCommand: vi.fn().mockReturnValue({ dispose: vi.fn() }),
-}))
-
 describe("getVisibleProviderOrLog", () => {
 	let mockOutputChannel: vscode.OutputChannel
 
@@ -176,14 +172,6 @@ describe("registerCommands handlers", () => {
 		setPanel(undefined, "tab")
 	})
 
-	it("registers the ripgrep diagnostic command and stores its disposable in context.subscriptions", async () => {
-		const { registerRipgrepDiagnosticCommand } = await import("../../services/ripgrep/diagnostic")
-		const mock = vi.mocked(registerRipgrepDiagnosticCommand)
-		const disposable = mock.mock.results[0]?.value
-		expect(mock).toHaveBeenCalled()
-		expect(mockContext.subscriptions).toContain(disposable)
-	})
-
 	it("settingsButtonClicked posts both settingsButtonClicked and didBecomeVisible actions", () => {
 		handlers["zoo-code.settingsButtonClicked"]()
 

src/activate/registerCommands.ts

@@ -13,7 +13,6 @@ import { handleNewTask } from "./handleTask"
 import { CodeIndexManager } from "../services/code-index/manager"
 import { importSettingsWithFeedback } from "../core/config/importExport"
 import { MdmService } from "../services/mdm/MdmService"
-import { registerRipgrepDiagnosticCommand } from "../services/ripgrep/diagnostic"
 import { t } from "../i18n"
 
 /**
@@ -69,27 +68,9 @@ export const registerCommands = (options: RegisterCommandOptions) => {
 		const command = getCommand(id as CommandId)
 		context.subscriptions.push(vscode.commands.registerCommand(command, callback))
 	}
-
-	context.subscriptions.push(registerRipgrepDiagnosticCommand())
 }
 
-// `showRipgrepDiagnostic` is registered separately by
-// `registerRipgrepDiagnosticCommand` (above), which owns the OutputChannel
-// lifecycle alongside the command registration, so it's intentionally
-// excluded from this map.
-//
-// Callback shape mirrors VS Code's own `commands.registerCommand` signature
-// (`(...args: any[]) => any`), with the return narrowed to `unknown` so
-// callers must inspect before using. `any[]` for args is unavoidable: the
-// callbacks here are heterogeneous (`importSettings` takes an optional
-// `filePath?: string`, others take none) and VS Code dispatches positional
-// args dynamically.
-type CommandCallback = (...args: any[]) => unknown
-const getCommandsMap = ({
-	context,
-	outputChannel,
-	provider,
-}: RegisterCommandOptions): Record<Exclude<CommandId, "showRipgrepDiagnostic">, CommandCallback> => ({
+const getCommandsMap = ({ context, outputChannel, provider }: RegisterCommandOptions): Record<CommandId, any> => ({
 	activationCompleted: () => {},
 	plusButtonClicked: async () => {
 		const visibleProvider = getVisibleProviderOrLog(outputChannel)

src/api/index.ts

@@ -90,6 +90,12 @@ export interface ApiHandlerCreateMessageMetadata {
 	 * Only applies to providers that support function calling restrictions (e.g., Gemini).
 	 */
 	allowedFunctionNames?: string[]
+	/**
+	 * Abort signal for cancelling the HTTP request mid-stream.
+	 * Passed through to AI SDK's streamText() so the underlying HTTP request is aborted
+	 * when the user clicks stop, preventing wasted API tokens/compute on the provider side.
+	 */
+	abortSignal?: AbortSignal
 }
 
 export interface ApiHandler {

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

@@ -354,7 +354,7 @@ describe("BaseOpenAiCompatibleProvider", () => {
 					stream: true,
 					stream_options: { include_usage: true },
 				}),
-				undefined,
+				expect.any(Object),
 			)
 		})
 

src/api/providers/__tests__/deepseek.spec.ts

@@ -637,6 +637,48 @@ describe("DeepSeekHandler", () => {
 			const toolCallChunks = chunks.filter((chunk) => chunk.type === "tool_call_partial")
 			expect(toolCallChunks.length).toBeGreaterThan(0)
 			expect(toolCallChunks[0].name).toBe("get_weather")
+
+			describe("abortSignal support", () => {
+				it("should pass abortSignal to chat.completions.create when provided in metadata", async () => {
+					const handler = new DeepSeekHandler({ ...mockOptions, apiKey: "test-key" })
+					const systemPrompt = "You are a helpful assistant."
+					const messages: Anthropic.Messages.MessageParam[] = [
+						{ role: "user", content: [{ type: "text" as const, text: "Hello!" }] },
+					]
+
+					const controller = new AbortController()
+					const mockAbortSignal = controller.signal
+
+					await handler.createMessage(systemPrompt, messages, {
+						taskId: "test",
+						abortSignal: mockAbortSignal,
+					})
+					for await (const _chunk of handler.createMessage(systemPrompt, messages)) {
+						break
+					}
+
+					expect(mockCreate).toHaveBeenCalled()
+					const callArgs = mockCreate.mock.calls[0][0]
+					expect(callArgs.signal).toBe(mockAbortSignal)
+				})
+
+				it("should not include signal when abortSignal is not provided", async () => {
+					const handler = new DeepSeekHandler({ ...mockOptions, apiKey: "test-key" })
+					const systemPrompt = "You are a helpful assistant."
+					const messages: Anthropic.Messages.MessageParam[] = [
+						{ role: "user", content: [{ type: "text" as const, text: "Hello!" }] },
+					]
+
+					await handler.createMessage(systemPrompt, messages)
+					for await (const _chunk of handler.createMessage(systemPrompt, messages)) {
+						break
+					}
+
+					expect(mockCreate).toHaveBeenCalled()
+					const callArgs = mockCreate.mock.calls[0][0]
+					expect(callArgs.signal).toBeUndefined()
+				})
+			})
 		})
 	})
 })

src/api/providers/__tests__/fireworks.spec.ts

@@ -95,25 +95,46 @@ describe("FireworksHandler", () => {
 	})
 
 	it.each([
-		{ modelId: "accounts/fireworks/models/glm-5p1" as const, contextWindow: 202752, inputPrice: 1.4, outputPrice: 4.4, cacheReadsPrice: 0.26 },
-		{ modelId: "accounts/fireworks/models/kimi-k2p6" as const, contextWindow: 262144, inputPrice: 0.95, outputPrice: 4.0, cacheReadsPrice: 0.16 },
-		{ modelId: "accounts/fireworks/models/deepseek-v4-pro" as const, contextWindow: 1048576, inputPrice: 1.74, outputPrice: 3.48, cacheReadsPrice: 0.14 },
-	])("should expose newly added model $modelId", ({ modelId, contextWindow, inputPrice, outputPrice, cacheReadsPrice }) => {
-		expect(fireworksModels[modelId]).toBeDefined()
-		const info = fireworksModels[modelId]
-		expect(info.maxTokens).toBeGreaterThan(0)
-		expect(info.contextWindow).toBe(contextWindow)
-		expect(info.inputPrice).toBe(inputPrice)
-		expect(info.outputPrice).toBe(outputPrice)
-		expect(info.cacheReadsPrice).toBe(cacheReadsPrice)
-		expect(info.description).toBeTruthy()
-
-		const handlerWithModel = new FireworksHandler({
-			apiModelId: modelId,
-			fireworksApiKey: "test-fireworks-api-key",
-		})
-		expect(handlerWithModel.getModel().id).toBe(modelId)
-	})
+		{
+			modelId: "accounts/fireworks/models/glm-5p1" as const,
+			contextWindow: 202752,
+			inputPrice: 1.4,
+			outputPrice: 4.4,
+			cacheReadsPrice: 0.26,
+		},
+		{
+			modelId: "accounts/fireworks/models/kimi-k2p6" as const,
+			contextWindow: 262144,
+			inputPrice: 0.95,
+			outputPrice: 4.0,
+			cacheReadsPrice: 0.16,
+		},
+		{
+			modelId: "accounts/fireworks/models/deepseek-v4-pro" as const,
+			contextWindow: 1048576,
+			inputPrice: 1.74,
+			outputPrice: 3.48,
+			cacheReadsPrice: 0.14,
+		},
+	])(
+		"should expose newly added model $modelId",
+		({ modelId, contextWindow, inputPrice, outputPrice, cacheReadsPrice }) => {
+			expect(fireworksModels[modelId]).toBeDefined()
+			const info = fireworksModels[modelId]
+			expect(info.maxTokens).toBeGreaterThan(0)
+			expect(info.contextWindow).toBe(contextWindow)
+			expect(info.inputPrice).toBe(inputPrice)
+			expect(info.outputPrice).toBe(outputPrice)
+			expect(info.cacheReadsPrice).toBe(cacheReadsPrice)
+			expect(info.description).toBeTruthy()
+
+			const handlerWithModel = new FireworksHandler({
+				apiModelId: modelId,
+				fireworksApiKey: "test-fireworks-api-key",
+			})
+			expect(handlerWithModel.getModel().id).toBe(modelId)
+		},
+	)
 
 	it("should return Kimi K2 Instruct model with correct configuration", () => {
 		const testModelId: FireworksModelId = "accounts/fireworks/models/kimi-k2-instruct"
@@ -465,7 +486,7 @@ describe("FireworksHandler", () => {
 				stream: true,
 				stream_options: { include_usage: true },
 			}),
-			undefined,
+			expect.any(Object),
 		)
 	})
 
@@ -491,7 +512,7 @@ describe("FireworksHandler", () => {
 			expect.objectContaining({
 				temperature: 0.5,
 			}),
-			undefined,
+			expect.any(Object),
 		)
 	})
 
@@ -518,7 +539,7 @@ describe("FireworksHandler", () => {
 			expect.objectContaining({
 				temperature: 1.0,
 			}),
-			undefined,
+			expect.any(Object),
 		)
 	})
 
@@ -546,7 +567,7 @@ describe("FireworksHandler", () => {
 			expect.objectContaining({
 				temperature: 0.7,
 			}),
-			undefined,
+			expect.any(Object),
 		)
 	})
 

src/api/providers/__tests__/lmstudio-native-tools.spec.ts

@@ -81,6 +81,7 @@ describe("LmStudioHandler Native Tools", () => {
 						}),
 					]),
 				}),
+				expect.any(Object),
 			)
 			// parallel_tool_calls should be true by default when not explicitly set
 			const callArgs = mockCreate.mock.calls[0][0]
@@ -107,6 +108,7 @@ describe("LmStudioHandler Native Tools", () => {
 				expect.objectContaining({
 					tool_choice: "auto",
 				}),
+				expect.any(Object),
 			)
 		})
 
@@ -219,6 +221,7 @@ describe("LmStudioHandler Native Tools", () => {
 				expect.objectContaining({
 					parallel_tool_calls: true,
 				}),
+				expect.any(Object),
 			)
 		})
 

src/api/providers/__tests__/mimo.spec.ts

@@ -374,6 +374,7 @@ describe("MimoHandler", () => {
 				expect.objectContaining({
 					extra_body: { thinking: { type: "enabled" } },
 				}),
+				expect.any(Object),
 			)
 		})