style: 完成所有文件的lint

2026-06-21 07:45:52 +00:00 · 2026-05-01 21:39:30 +08:00
parent d136872cc9
commit 6182015005
1333 changed files with 68255 additions and 77882 deletions
--- a/src/services/api/claude.ts
+++ b/src/services/api/claude.ts
@@ -230,7 +230,11 @@ import { getInitializationStatus } from '../lsp/manager.js'
 import { isToolFromMcpServer } from '../mcp/utils.js'
 import { recordLLMObservation } from '../langfuse/index.js'
 import type { LangfuseSpan } from '../langfuse/index.js'
-import { convertMessagesToLangfuse, convertOutputToLangfuse, convertToolsToLangfuse } from '../langfuse/convert.js'
+import {
+  convertMessagesToLangfuse,
+  convertOutputToLangfuse,
+  convertToolsToLangfuse,
+} from '../langfuse/convert.js'
 import { withStreamingVCR, withVCR } from '../vcr.js'
 import { CLIENT_REQUEST_ID_HEADER, getAnthropicClient } from './client.js'
 import {
@@ -442,7 +446,7 @@ function configureEffortParams(
    betas.push(EFFORT_BETA_HEADER)
  } else if (typeof effortValue === 'string') {
    // Send string effort level as is
-    outputConfig.effort = effortValue as "high" | "medium" | "low" | "max"
+    outputConfig.effort = effortValue as 'high' | 'medium' | 'low' | 'max'
    betas.push(EFFORT_BETA_HEADER)
  } else if (process.env.USER_TYPE === 'ant') {
    // Numeric effort override - ant-only (uses anthropic_internal)
@@ -615,7 +619,8 @@ export function userMessageToMessageParam(
    role: 'user',
    content: (Array.isArray(message.message!.content)
      ? [...message.message!.content]
-      : message.message!.content) as import('@anthropic-ai/sdk/resources/beta/messages/messages.js').BetaContentBlockParam[],
+      : message.message!
+          .content) as import('@anthropic-ai/sdk/resources/beta/messages/messages.js').BetaContentBlockParam[],
  }
 }

@@ -666,7 +671,9 @@ export function assistantMessageToMessageParam(
    content:
      typeof message.message!.content === 'string'
        ? message.message!.content
-        : message.message!.content!.map(stripGeminiProviderMetadata) as BetaContentBlockParam[],
+        : (message.message!.content!.map(
+            stripGeminiProviderMetadata,
+          ) as BetaContentBlockParam[]),
  }
 }

@@ -681,10 +688,8 @@ function stripGeminiProviderMetadata<T extends BetaContentBlockParam | string>(
  }

  const obj = contentBlock as unknown as Record<string, unknown>
-  const {
-    _geminiThoughtSignature: _unusedGeminiThoughtSignature,
-    ...rest
-  } = obj
+  const { _geminiThoughtSignature: _unusedGeminiThoughtSignature, ...rest } =
+    obj
  return rest as unknown as T
 }

@@ -1343,7 +1348,13 @@ async function* queryModel(
    // OpenAI emulates Anthropic's dynamic tool loading client-side. It needs
    // the full tool pool so ToolSearchTool can search deferred MCP tools that
    // were intentionally filtered out of the initial API tool list above.
-    yield* queryModelOpenAI(messagesForAPI, systemPrompt, tools, signal, options)
+    yield* queryModelOpenAI(
+      messagesForAPI,
+      systemPrompt,
+      tools,
+      signal,
+      options,
+    )
    return
  }

@@ -1362,7 +1373,13 @@ async function* queryModel(

  if (getAPIProvider() === 'grok') {
    const { queryModelGrok } = await import('./grok/index.js')
-    yield* queryModelGrok(messagesForAPI, systemPrompt, filteredTools, signal, options)
+    yield* queryModelGrok(
+      messagesForAPI,
+      systemPrompt,
+      filteredTools,
+      signal,
+      options,
+    )
    return
  }

@@ -2127,7 +2144,8 @@ async function* queryModel(
                })
                throw new Error('Content block is not a connector_text block')
              }
-              ;(contentBlock as { connector_text: string }).connector_text += delta.connector_text
+              ;(contentBlock as { connector_text: string }).connector_text +=
+                delta.connector_text
            } else {
              switch (delta.type) {
                case 'citations_delta':
@@ -2206,7 +2224,8 @@ async function* queryModel(
                    })
                    throw new Error('Content block is not a thinking block')
                  }
-                  ;(contentBlock as { thinking: string }).thinking += delta.thinking
+                  ;(contentBlock as { thinking: string }).thinking +=
+                    delta.thinking
                  break
              }
            }
@@ -2298,7 +2317,10 @@ async function* queryModel(
            }

            // Update cost
-            const costUSDForPart = calculateUSDCost(resolvedModel, usage as unknown as BetaUsage)
+            const costUSDForPart = calculateUSDCost(
+              resolvedModel,
+              usage as unknown as BetaUsage,
+            )
            costUSD += addToTotalSessionCost(
              costUSDForPart,
              usage as unknown as BetaUsage,
@@ -2887,10 +2909,14 @@ async function* queryModel(
    // message_delta handler before any yield. Fallback pushes to newMessages
    // then yields, so tracking must be here to survive .return() at the yield.
    if (fallbackMessage) {
-      const fallbackUsage = fallbackMessage.message.usage as BetaMessageDeltaUsage
+      const fallbackUsage = fallbackMessage.message
+        .usage as BetaMessageDeltaUsage
      usage = updateUsage(EMPTY_USAGE, fallbackUsage)
      stopReason = fallbackMessage.message.stop_reason as BetaStopReason
-      const fallbackCost = calculateUSDCost(resolvedModel, fallbackUsage as unknown as BetaUsage)
+      const fallbackCost = calculateUSDCost(
+        resolvedModel,
+        fallbackUsage as unknown as BetaUsage,
+      )
      costUSD += addToTotalSessionCost(
        fallbackCost,
        fallbackUsage as unknown as BetaUsage,
@@ -2946,7 +2972,9 @@ async function* queryModel(
  void options.getToolPermissionContext().then(permissionContext => {
    logAPISuccessAndDuration({
      model:
-        (newMessages[0]?.message.model as string | undefined) ?? partialMessage?.model ?? options.model,
+        (newMessages[0]?.message.model as string | undefined) ??
+        partialMessage?.model ??
+        options.model,
      preNormalizedModel: options.model,
      usage,
      start,
--- a/src/services/api/gemini/index.ts
+++ b/src/services/api/gemini/index.ts
@@ -19,9 +19,20 @@ import type { SystemPrompt } from '../../../utils/systemPromptType.js'
 import type { ThinkingConfig } from '../../../utils/thinking.js'
 import type { Options } from '../claude.js'
 import { recordLLMObservation } from '../../../services/langfuse/tracing.js'
-import { convertMessagesToLangfuse, convertOutputToLangfuse, convertToolsToLangfuse } from '../../../services/langfuse/convert.js'
+import {
+  convertMessagesToLangfuse,
+  convertOutputToLangfuse,
+  convertToolsToLangfuse,
+} from '../../../services/langfuse/convert.js'
 import { streamGeminiGenerateContent } from './client.js'
-import { anthropicMessagesToGemini, resolveGeminiModel, adaptGeminiStreamToAnthropic, anthropicToolsToGemini, anthropicToolChoiceToGemini, GEMINI_THOUGHT_SIGNATURE_FIELD } from '@ant/model-provider'
+import {
+  anthropicMessagesToGemini,
+  resolveGeminiModel,
+  adaptGeminiStreamToAnthropic,
+  anthropicToolsToGemini,
+  anthropicToolChoiceToGemini,
+  GEMINI_THOUGHT_SIGNATURE_FIELD,
+} from '@ant/model-provider'

 export async function* queryModelGemini(
  messages: Message[],
@@ -209,7 +220,9 @@ export async function* queryModelGemini(
    yield createAssistantAPIErrorMessage({
      content: `API Error: ${errorMessage}`,
      apiError: 'api_error',
-      error: (error instanceof Error ? error : new Error(String(error))) as unknown as SDKAssistantMessageError,
+      error: (error instanceof Error
+        ? error
+        : new Error(String(error))) as unknown as SDKAssistantMessageError,
    })
  }
 }
--- a/src/services/api/grok/tests/client.test.ts
+++ b/src/services/api/grok/tests/client.test.ts
@@ -2,7 +2,7 @@ import { describe, expect, test, beforeEach, afterEach, mock } from 'bun:test'

 // Defensive: agent.test.ts can corrupt Bun's src/* path alias at runtime.
 mock.module('src/utils/proxy.js', () => ({
-  getProxyFetchOptions: () => ({} as any),
+  getProxyFetchOptions: () => ({}) as any,
 }))

 import { getGrokClient, clearGrokClientCache } from '../client.js'
--- a/src/services/api/grok/index.ts
+++ b/src/services/api/grok/index.ts
@@ -1,13 +1,24 @@
 import type { BetaToolUnion } from '@anthropic-ai/sdk/resources/beta/messages/messages.mjs'
 import type { SystemPrompt } from '../../../utils/systemPromptType.js'
-import type { Message, StreamEvent, SystemAPIErrorMessage, AssistantMessage } from '../../../types/message.js'
+import type {
+  Message,
+  StreamEvent,
+  SystemAPIErrorMessage,
+  AssistantMessage,
+} from '../../../types/message.js'
 import type { Tools } from '../../../Tool.js'
 import type {
  ChatCompletionChunk,
  ChatCompletionCreateParamsStreaming,
 } from 'openai/resources/chat/completions/completions.mjs'
 import { getGrokClient } from './client.js'
-import { anthropicMessagesToOpenAI, anthropicToolsToOpenAI, anthropicToolChoiceToOpenAI, adaptOpenAIStreamToAnthropic, resolveGrokModel } from '@ant/model-provider'
+import {
+  anthropicMessagesToOpenAI,
+  anthropicToolsToOpenAI,
+  anthropicToolChoiceToOpenAI,
+  adaptOpenAIStreamToAnthropic,
+  resolveGrokModel,
+} from '@ant/model-provider'
 import { normalizeMessagesForAPI } from '../../../utils/messages.js'
 import type { SDKAssistantMessageError } from '../../../entrypoints/agentSdkTypes.js'
 import { toolToAPISchema } from '../../../utils/api.js'
@@ -15,7 +26,11 @@ import { logForDebugging } from '../../../utils/debug.js'
 import { addToTotalSessionCost } from '../../../cost-tracker.js'
 import { calculateUSDCost } from '../../../utils/modelCost.js'
 import { recordLLMObservation } from '../../../services/langfuse/tracing.js'
-import { convertMessagesToLangfuse, convertOutputToLangfuse, convertToolsToLangfuse } from '../../../services/langfuse/convert.js'
+import {
+  convertMessagesToLangfuse,
+  convertOutputToLangfuse,
+  convertToolsToLangfuse,
+} from '../../../services/langfuse/convert.js'
 import type { Options } from '../claude.js'
 import { randomUUID } from 'crypto'
 import {
@@ -56,11 +71,16 @@ export async function* queryModelGrok(
    const standardTools = toolSchemas.filter(
      (t): t is BetaToolUnion & { type: string } => {
        const anyT = t as unknown as Record<string, unknown>
-        return anyT.type !== 'advisor_20260301' && anyT.type !== 'computer_20250124'
+        return (
+          anyT.type !== 'advisor_20260301' && anyT.type !== 'computer_20250124'
+        )
      },
    )

-    const openaiMessages = anthropicMessagesToOpenAI(messagesForAPI, systemPrompt)
+    const openaiMessages = anthropicMessagesToOpenAI(
+      messagesForAPI,
+      systemPrompt,
+    )
    const openaiTools = anthropicToolsToOpenAI(standardTools)
    const openaiToolChoice = anthropicToolChoiceToOpenAI(options.toolChoice)

@@ -70,7 +90,9 @@ export async function* queryModelGrok(
      source: options.querySource,
    })

-    logForDebugging(`[Grok] Calling model=${grokModel}, messages=${openaiMessages.length}, tools=${openaiTools.length}`)
+    logForDebugging(
+      `[Grok] Calling model=${grokModel}, messages=${openaiMessages.length}, tools=${openaiTools.length}`,
+    )

    const stream = await client.chat.completions.create(
      {
@@ -91,7 +113,10 @@ export async function* queryModelGrok(
      },
    )

-    const adaptedStream = adaptOpenAIStreamToAnthropic(stream as AsyncIterable<ChatCompletionChunk>, grokModel)
+    const adaptedStream = adaptOpenAIStreamToAnthropic(
+      stream as AsyncIterable<ChatCompletionChunk>,
+      grokModel,
+    )

    const contentBlocks: Record<number, any> = {}
    const collectedMessages: AssistantMessage[] = []
@@ -111,7 +136,7 @@ export async function* queryModelGrok(
          partialMessage = (event as any).message
          ttftMs = Date.now() - start
          if ((event as any).message?.usage) {
-            usage = { ...usage, ...((event as any).message.usage) }
+            usage = { ...usage, ...(event as any).message.usage }
          }
          break
        }
@@ -175,7 +200,10 @@ export async function* queryModelGrok(
          break
      }

-      if (event.type === 'message_stop' && usage.input_tokens + usage.output_tokens > 0) {
+      if (
+        event.type === 'message_stop' &&
+        usage.input_tokens + usage.output_tokens > 0
+      ) {
        const costUSD = calculateUSDCost(grokModel, usage as any)
        addToTotalSessionCost(costUSD, usage as any, options.model)
      }
@@ -210,7 +238,9 @@ export async function* queryModelGrok(
    yield createAssistantAPIErrorMessage({
      content: `API Error: ${errorMessage}`,
      apiError: 'api_error',
-      error: (error instanceof Error ? error : new Error(String(error))) as unknown as SDKAssistantMessageError,
+      error: (error instanceof Error
+        ? error
+        : new Error(String(error))) as unknown as SDKAssistantMessageError,
    })
  }
 }
--- a/src/services/api/logging.ts
+++ b/src/services/api/logging.ts
@@ -195,7 +195,8 @@ export function logAPIQuery({
  previousRequestId?: string | null
 }): void {
  const thinkingType = thinkingConfig?.type ?? 'disabled'
-  const thinkingBudgetTokens = thinkingConfig?.type === 'enabled' ? thinkingConfig.budgetTokens : undefined
+  const thinkingBudgetTokens =
+    thinkingConfig?.type === 'enabled' ? thinkingConfig.budgetTokens : undefined
  logEvent('tengu_api_query', {
    model: model as AnalyticsMetadata_I_VERIFIED_THIS_IS_NOT_CODE_OR_FILEPATHS,
    messagesLength,
@@ -662,7 +663,9 @@ export function logAPISuccessAndDuration({
    let connectorCount = 0

    for (const msg of newMessages) {
-      const contentArr = Array.isArray(msg.message.content) ? msg.message.content : []
+      const contentArr = Array.isArray(msg.message.content)
+        ? msg.message.content
+        : []
      for (const block of contentArr) {
        if (typeof block === 'string') continue
        if (block.type === 'text') {
@@ -670,14 +673,19 @@ export function logAPISuccessAndDuration({
        } else if (feature('CONNECTOR_TEXT') && isConnectorTextBlock(block)) {
          connectorCount++
        } else if (block.type === 'thinking') {
-          thinkingLen += (block as { type: 'thinking'; thinking: string }).thinking.length
+          thinkingLen += (block as { type: 'thinking'; thinking: string })
+            .thinking.length
        } else if (
          block.type === 'tool_use' ||
          block.type === 'server_tool_use' ||
          (block.type as string) === 'mcp_tool_use'
        ) {
-          const inputLen = jsonStringify((block as { input: unknown }).input).length
-          const sanitizedName = sanitizeToolNameForAnalytics((block as { name: string }).name)
+          const inputLen = jsonStringify(
+            (block as { input: unknown }).input,
+          ).length
+          const sanitizedName = sanitizeToolNameForAnalytics(
+            (block as { name: string }).name,
+          )
          toolLengths[sanitizedName] =
            (toolLengths[sanitizedName] ?? 0) + inputLen
          hasToolUse = true
--- a/src/services/api/openai/tests/thinking.test.ts
+++ b/src/services/api/openai/tests/thinking.test.ts
@@ -1,5 +1,8 @@
 import { describe, expect, test, beforeEach, afterEach } from 'bun:test'
-import { isOpenAIThinkingEnabled, buildOpenAIRequestBody } from '../requestBody.js'
+import {
+  isOpenAIThinkingEnabled,
+  buildOpenAIRequestBody,
+} from '../requestBody.js'

 describe('isOpenAIThinkingEnabled', () => {
  const originalEnv = {
@@ -81,7 +84,9 @@ describe('isOpenAIThinkingEnabled', () => {
    })

    test('returns true when model name is namespaced for deepseek-reasoner', () => {
-      expect(isOpenAIThinkingEnabled('TokenService/deepseek-reasoner')).toBe(true)
+      expect(isOpenAIThinkingEnabled('TokenService/deepseek-reasoner')).toBe(
+        true,
+      )
    })

    test('returns true when model name is "deepseek-v3.2"', () => {
@@ -185,14 +190,20 @@ describe('buildOpenAIRequestBody — thinking params', () => {
  })

  test('does NOT include thinking params when disabled', () => {
-    const body = buildOpenAIRequestBody({ ...baseParams, enableThinking: false })
+    const body = buildOpenAIRequestBody({
+      ...baseParams,
+      enableThinking: false,
+    })
    expect(body.thinking).toBeUndefined()
    expect(body.enable_thinking).toBeUndefined()
    expect(body.chat_template_kwargs).toBeUndefined()
  })

  test('always includes stream and stream_options', () => {
-    const body = buildOpenAIRequestBody({ ...baseParams, enableThinking: false })
+    const body = buildOpenAIRequestBody({
+      ...baseParams,
+      enableThinking: false,
+    })
    expect(body.stream).toBe(true)
    expect(body.stream_options).toEqual({ include_usage: true })
  })
@@ -216,7 +227,10 @@ describe('buildOpenAIRequestBody — thinking params', () => {
  })

  test('excludes temperature when thinking is off and no override', () => {
-    const body = buildOpenAIRequestBody({ ...baseParams, enableThinking: false })
+    const body = buildOpenAIRequestBody({
+      ...baseParams,
+      enableThinking: false,
+    })
    expect(body.temperature).toBeUndefined()
  })

@@ -232,8 +246,11 @@ describe('buildOpenAIRequestBody — thinking params', () => {
  })

  test('excludes tools when empty', () => {
-    const body = buildOpenAIRequestBody({ ...baseParams, enableThinking: false })
+    const body = buildOpenAIRequestBody({
+      ...baseParams,
+      enableThinking: false,
+    })
    expect(body.tools).toBeUndefined()
    expect(body.tool_choice).toBeUndefined()
  })
-})
+})
--- a/src/services/api/openai/client.ts
+++ b/src/services/api/openai/client.ts
@@ -56,8 +56,12 @@ export function getOpenAIClient(options?: {
    maxRetries: options?.maxRetries ?? 0,
    timeout: parseInt(process.env.API_TIMEOUT_MS || String(600 * 1000), 10),
    dangerouslyAllowBrowser: true,
-    ...(process.env.OPENAI_ORG_ID && { organization: process.env.OPENAI_ORG_ID }),
-    ...(process.env.OPENAI_PROJECT_ID && { project: process.env.OPENAI_PROJECT_ID }),
+    ...(process.env.OPENAI_ORG_ID && {
+      organization: process.env.OPENAI_ORG_ID,
+    }),
+    ...(process.env.OPENAI_PROJECT_ID && {
+      project: process.env.OPENAI_PROJECT_ID,
+    }),
    fetchOptions: getProxyFetchOptions({ forAnthropicAPI: false }),
    fetch: wrappedFetch,
  })
--- a/src/services/api/openai/index.ts
+++ b/src/services/api/openai/index.ts
@@ -10,11 +10,15 @@ import type {
 import type { AgentId } from '../../../types/ids.js'
 import type { Tools } from '../../../Tool.js'
 import type { Stream } from 'openai/streaming.mjs'
-import type {
-  ChatCompletionCreateParamsStreaming,
-} from 'openai/resources/chat/completions/completions.mjs'
+import type { ChatCompletionCreateParamsStreaming } from 'openai/resources/chat/completions/completions.mjs'
 import { getOpenAIClient } from './client.js'
-import { anthropicMessagesToOpenAI, resolveOpenAIModel, adaptOpenAIStreamToAnthropic, anthropicToolsToOpenAI, anthropicToolChoiceToOpenAI } from '@ant/model-provider'
+import {
+  anthropicMessagesToOpenAI,
+  resolveOpenAIModel,
+  adaptOpenAIStreamToAnthropic,
+  anthropicToolsToOpenAI,
+  anthropicToolChoiceToOpenAI,
+} from '@ant/model-provider'
 import { normalizeMessagesForAPI } from '../../../utils/messages.js'
 import { toolToAPISchema } from '../../../utils/api.js'
 import {
@@ -24,10 +28,22 @@ import {
 import { logForDebugging } from '../../../utils/debug.js'
 import { addToTotalSessionCost } from '../../../cost-tracker.js'
 import { calculateUSDCost } from '../../../utils/modelCost.js'
-import { isOpenAIThinkingEnabled, resolveOpenAIMaxTokens, buildOpenAIRequestBody } from './requestBody.js'
+import {
+  isOpenAIThinkingEnabled,
+  resolveOpenAIMaxTokens,
+  buildOpenAIRequestBody,
+} from './requestBody.js'
 import { recordLLMObservation } from '../../../services/langfuse/tracing.js'
-import { convertMessagesToLangfuse, convertOutputToLangfuse, convertToolsToLangfuse } from '../../../services/langfuse/convert.js'
-export { isOpenAIThinkingEnabled, resolveOpenAIMaxTokens, buildOpenAIRequestBody }
+import {
+  convertMessagesToLangfuse,
+  convertOutputToLangfuse,
+  convertToolsToLangfuse,
+} from '../../../services/langfuse/convert.js'
+export {
+  isOpenAIThinkingEnabled,
+  resolveOpenAIMaxTokens,
+  buildOpenAIRequestBody,
+}
 import { getModelMaxOutputTokens } from '../../../utils/context.js'
 import type { Options } from '../claude.js'
 import { randomUUID } from 'crypto'
@@ -81,7 +97,9 @@ function prependDeferredToolListIfNeeded(
  ]
 }

-function isOpenAIConvertibleMessage(msg: Message): msg is AssistantMessage | UserMessage {
+function isOpenAIConvertibleMessage(
+  msg: Message,
+): msg is AssistantMessage | UserMessage {
  return msg.type === 'assistant' || msg.type === 'user'
 }

@@ -95,11 +113,24 @@ function assembleFinalAssistantOutputs(params: {
  contentBlocks: Record<number, any>
  tools: Tools
  agentId: string | undefined
-  usage: { input_tokens: number; output_tokens: number; cache_creation_input_tokens: number; cache_read_input_tokens: number }
+  usage: {
+    input_tokens: number
+    output_tokens: number
+    cache_creation_input_tokens: number
+    cache_read_input_tokens: number
+  }
  stopReason: string | null
  maxTokens: number
 }): (AssistantMessage | SystemAPIErrorMessage)[] {
-  const { partialMessage, contentBlocks, tools, agentId, usage, stopReason, maxTokens } = params
+  const {
+    partialMessage,
+    contentBlocks,
+    tools,
+    agentId,
+    usage,
+    stopReason,
+    maxTokens,
+  } = params
  const outputs: (AssistantMessage | SystemAPIErrorMessage)[] = []

  const allBlocks = Object.keys(contentBlocks)
@@ -111,7 +142,11 @@ function assembleFinalAssistantOutputs(params: {
    outputs.push({
      message: {
        ...partialMessage,
-        content: normalizeContentFromAPI(allBlocks, tools, agentId as AgentId | undefined),
+        content: normalizeContentFromAPI(
+          allBlocks,
+          tools,
+          agentId as AgentId | undefined,
+        ),
        usage,
        stop_reason: stopReason,
        stop_sequence: null,
@@ -124,12 +159,15 @@ function assembleFinalAssistantOutputs(params: {
  }

  if (stopReason === 'max_tokens') {
-    outputs.push(createAssistantAPIErrorMessage({
-      content: `Output truncated: response exceeded the ${maxTokens} token limit. ` +
-        `Set OPENAI_MAX_TOKENS or CLAUDE_CODE_MAX_OUTPUT_TOKENS to override.`,
-      apiError: 'max_output_tokens',
-      error: 'max_output_tokens',
-    }))
+    outputs.push(
+      createAssistantAPIErrorMessage({
+        content:
+          `Output truncated: response exceeded the ${maxTokens} token limit. ` +
+          `Set OPENAI_MAX_TOKENS or CLAUDE_CODE_MAX_OUTPUT_TOKENS to override.`,
+        apiError: 'max_output_tokens',
+        error: 'max_output_tokens',
+      }),
+    )
  }

  return outputs
@@ -217,7 +255,9 @@ export async function* queryModelOpenAI(

    // 8. Convert messages and tools to OpenAI format
    const enableThinking = isOpenAIThinkingEnabled(openaiModel)
-    const openAIConvertibleMessages = messagesForAPI.filter(isOpenAIConvertibleMessage)
+    const openAIConvertibleMessages = messagesForAPI.filter(
+      isOpenAIConvertibleMessage,
+    )
    const messagesWithDeferredToolList = prependDeferredToolListIfNeeded(
      openAIConvertibleMessages,
      tools,
@@ -264,7 +304,10 @@ export async function* queryModelOpenAI(
    //     3. CLAUDE_CODE_MAX_OUTPUT_TOKENS env var (generic override)
    //     4. upperLimit default (64000)
    const { upperLimit } = getModelMaxOutputTokens(openaiModel)
-    const maxTokens = resolveOpenAIMaxTokens(upperLimit, options.maxOutputTokensOverride)
+    const maxTokens = resolveOpenAIMaxTokens(
+      upperLimit,
+      options.maxOutputTokensOverride,
+    )

    // 11. Get client
    const client = getOpenAIClient({
@@ -287,10 +330,7 @@ export async function* queryModelOpenAI(
      maxTokens,
      temperatureOverride: options.temperatureOverride,
    })
-    const stream = await client.chat.completions.create(
-      requestBody,
-      { signal },
-    )
+    const stream = await client.chat.completions.create(requestBody, { signal })

    // 12. Convert OpenAI stream to Anthropic events, then process into
    //     AssistantMessage + StreamEvent (matching the Anthropic path behavior)
@@ -373,8 +413,13 @@ export async function* queryModelOpenAI(
          // here and injected so tokenCountWithEstimation() can read it.
          if (partialMessage) {
            for (const output of assembleFinalAssistantOutputs({
-              partialMessage, contentBlocks, tools, agentId: options.agentId,
-              usage, stopReason, maxTokens,
+              partialMessage,
+              contentBlocks,
+              tools,
+              agentId: options.agentId,
+              usage,
+              stopReason,
+              maxTokens,
            })) {
              if (output.type === 'assistant') {
                collectedMessages.push(output)
@@ -424,8 +469,13 @@ export async function* queryModelOpenAI(
    // Safety: if stream ended without message_stop, assemble and yield whatever we have
    if (partialMessage) {
      for (const output of assembleFinalAssistantOutputs({
-        partialMessage, contentBlocks, tools, agentId: options.agentId,
-        usage, stopReason, maxTokens,
+        partialMessage,
+        contentBlocks,
+        tools,
+        agentId: options.agentId,
+        usage,
+        stopReason,
+        maxTokens,
      })) {
        yield output
      }
@@ -436,7 +486,9 @@ export async function* queryModelOpenAI(
    yield createAssistantAPIErrorMessage({
      content: `API Error: ${errorMessage}`,
      apiError: 'api_error',
-      error: (error instanceof Error ? error : new Error(String(error))) as unknown as SDKAssistantMessageError,
+      error: (error instanceof Error
+        ? error
+        : new Error(String(error))) as unknown as SDKAssistantMessageError,
    })
  }
 }
--- a/src/services/api/openai/requestBody.ts
+++ b/src/services/api/openai/requestBody.ts
@@ -3,9 +3,7 @@
 * thinking mode. Extracted from index.ts so tests can import them without
 * triggering heavy module side-effects (OpenAI client, stream adapter, etc.).
 */
-import type {
-  ChatCompletionCreateParamsStreaming,
-} from 'openai/resources/chat/completions/completions.mjs'
+import type { ChatCompletionCreateParamsStreaming } from 'openai/resources/chat/completions/completions.mjs'
 import { isEnvTruthy, isEnvDefinedFalsy } from '../../../utils/envUtils.js'

 /**
@@ -44,10 +42,16 @@ export function resolveOpenAIMaxTokens(
  upperLimit: number,
  maxOutputTokensOverride?: number,
 ): number {
-  return maxOutputTokensOverride
-    ?? (process.env.OPENAI_MAX_TOKENS ? parseInt(process.env.OPENAI_MAX_TOKENS, 10) || undefined : undefined)
-    ?? (process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS ? parseInt(process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS, 10) || undefined : undefined)
-    ?? upperLimit
+  return (
+    maxOutputTokensOverride ??
+    (process.env.OPENAI_MAX_TOKENS
+      ? parseInt(process.env.OPENAI_MAX_TOKENS, 10) || undefined
+      : undefined) ??
+    (process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS
+      ? parseInt(process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS, 10) || undefined
+      : undefined) ??
+    upperLimit
+  )
 }

 /**
@@ -74,7 +78,15 @@ export function buildOpenAIRequestBody(params: {
  enable_thinking?: boolean
  chat_template_kwargs?: { thinking: boolean }
 } {
-  const { model, messages, tools, toolChoice, enableThinking, maxTokens, temperatureOverride } = params
+  const {
+    model,
+    messages,
+    tools,
+    toolChoice,
+    enableThinking,
+    maxTokens,
+    temperatureOverride,
+  } = params
  return {
    model,
    messages,
@@ -96,8 +108,9 @@ export function buildOpenAIRequestBody(params: {
    }),
    // Only send temperature when thinking mode is off (DeepSeek ignores it anyway,
    // but other providers may respect it)
-    ...(!enableThinking && temperatureOverride !== undefined && {
-      temperature: temperatureOverride,
-    }),
+    ...(!enableThinking &&
+      temperatureOverride !== undefined && {
+        temperature: temperatureOverride,
+      }),
  }
 }
--- a/src/services/api/promptCacheBreakDetection.ts
+++ b/src/services/api/promptCacheBreakDetection.ts
@@ -459,7 +459,8 @@ export async function checkResponseForCacheBreak(
    // assistant message timestamp in the messages array (before the current response)
    const lastAssistantMessage = messages.findLast(m => m.type === 'assistant')
    const timeSinceLastAssistantMsg = lastAssistantMessage
-      ? Date.now() - new Date(lastAssistantMessage.timestamp as string | number).getTime()
+      ? Date.now() -
+        new Date(lastAssistantMessage.timestamp as string | number).getTime()
      : null

    // Skip the first call — no previous value to compare against
--- a/src/services/api/withRetry.ts
+++ b/src/services/api/withRetry.ts
@@ -539,10 +539,7 @@ export function getRetryDelay(
    }
  }

-  const baseDelay = Math.min(
-    BASE_DELAY_MS * 2 ** (attempt - 1),
-    maxDelayMs,
-  )
+  const baseDelay = Math.min(BASE_DELAY_MS * 2 ** (attempt - 1), maxDelayMs)
  const jitter = Math.random() * 0.25 * baseDelay
  return baseDelay + jitter
 }