fix: address review feedback — enforce maxRetries, wire abort signal, use cross-platform shell

1. Enforce maxRetries: track auto-fix attempts per query chain in toolHooks.ts
   and stop feeding errors back after the configured limit is reached.

2. Wire abort signal to subprocess: subscribe to AbortController signal in
   runCommand() and kill the process tree on abort. Uses detached process
   groups on Unix to ensure child processes are also terminated.

3. Replace hardcoded bash with shell:true: use Node's cross-platform shell
   resolution instead of spawn('bash', ['-c', ...]) so auto-fix commands
   work on Windows and non-bash environments.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

.github/workflows/pr-checks.yml

@@ -29,6 +29,13 @@ jobs:
         with:
           bun-version: 1.3.11
 
+      - name: Set up Python
+        uses: actions/setup-python@0a5c61591373683505ea898e09a3ea4f39ef2b9c # v5.0.0
+        with:
+          python-version: "3.12"
+          cache: "pip"
+          cache-dependency-path: python/requirements.txt
+
       - name: Install dependencies
         run: bun install --frozen-lockfile
 
@@ -38,6 +45,12 @@ jobs:
       - name: Full unit test suite
         run: bun test --max-concurrency=1
 
+      - name: Install Python test dependencies
+        run: python -m pip install -r python/requirements.txt
+
+      - name: Python unit tests
+        run: python -m pytest -q python/tests
+
       - name: Suspicious PR intent scan
         run: bun run security:pr-scan -- --base ${{ github.event.pull_request.base.sha || 'origin/main' }}
       - name: Provider tests

python/requirements.txt

@@ -0,0 +1,3 @@
+pytest==7.4.4
+pytest-asyncio==0.23.3
+httpx==0.25.2

scripts/system-check.ts

@@ -118,14 +118,14 @@ function isLocalBaseUrl(baseUrl: string): boolean {
 }
 
 const GEMINI_DEFAULT_BASE_URL = 'https://generativelanguage.googleapis.com/v1beta/openai'
-const GITHUB_MODELS_DEFAULT_BASE = 'https://models.github.ai/inference'
+const GITHUB_COPILOT_BASE = 'https://api.githubcopilot.com'
 
 function currentBaseUrl(): string {
   if (isTruthy(process.env.CLAUDE_CODE_USE_GEMINI)) {
     return process.env.GEMINI_BASE_URL ?? GEMINI_DEFAULT_BASE_URL
   }
   if (isTruthy(process.env.CLAUDE_CODE_USE_GITHUB)) {
-    return process.env.OPENAI_BASE_URL ?? GITHUB_MODELS_DEFAULT_BASE
+    return process.env.OPENAI_BASE_URL ?? GITHUB_COPILOT_BASE
   }
   return process.env.OPENAI_BASE_URL ?? 'https://api.openai.com/v1'
 }
@@ -157,7 +157,7 @@ function checkGeminiEnv(): CheckResult[] {
 
 function checkGithubEnv(): CheckResult[] {
   const results: CheckResult[] = []
-  const baseUrl = process.env.OPENAI_BASE_URL ?? GITHUB_MODELS_DEFAULT_BASE
+  const baseUrl = process.env.OPENAI_BASE_URL ?? GITHUB_COPILOT_BASE
   results.push(pass('Provider mode', 'GitHub Models provider enabled.'))
 
   const token = process.env.GITHUB_TOKEN ?? process.env.GH_TOKEN
@@ -435,7 +435,7 @@ function serializeSafeEnvSummary(): Record<string, string | boolean> {
         process.env.OPENAI_MODEL ??
         '(unset, default: github:copilot → openai/gpt-4.1)',
       OPENAI_BASE_URL:
-        process.env.OPENAI_BASE_URL ?? GITHUB_MODELS_DEFAULT_BASE,
+        process.env.OPENAI_BASE_URL ?? GITHUB_COPILOT_BASE,
       GITHUB_TOKEN_SET: Boolean(
         process.env.GITHUB_TOKEN ?? process.env.GH_TOKEN,
       ),

src/commands.ts

@@ -136,6 +136,7 @@ import hooks from './commands/hooks/index.js'
 import files from './commands/files/index.js'
 import branch from './commands/branch/index.js'
 import agents from './commands/agents/index.js'
+import autoFix from './commands/auto-fix.js'
 import plugin from './commands/plugin/index.js'
 import reloadPlugins from './commands/reload-plugins/index.js'
 import rewind from './commands/rewind/index.js'
@@ -263,6 +264,7 @@ const COMMANDS = memoize((): Command[] => [
   addDir,
   advisor,
   agents,
+  autoFix,
   branch,
   btw,
   chrome,

src/commands/auto-fix.ts

@@ -0,0 +1,25 @@
+import type { Command } from '../types/command.js'
+
+const command: Command = {
+  name: 'auto-fix',
+  description: 'Configure auto-fix: run lint/test after AI edits',
+  isEnabled: () => true,
+  type: 'prompt',
+  progressMessage: 'Configuring auto-fix...',
+  contentLength: 0,
+  source: 'builtin',
+  async getPromptForCommand() {
+    return [
+      {
+        type: 'text',
+        text:
+          'The user wants to configure auto-fix settings. Auto-fix automatically runs lint and test commands after AI file edits, feeding errors back for self-repair.\n\n' +
+          'Current settings location: `.claude/settings.json` or `.claude/settings.local.json`\n\n' +
+          'Example configuration:\n```json\n{\n  "autoFix": {\n    "enabled": true,\n    "lint": "eslint . --fix",\n    "test": "bun test",\n    "maxRetries": 3,\n    "timeout": 30000\n  }\n}\n```\n\n' +
+          'Ask the user what lint and test commands they use, then help them set up the configuration.',
+      },
+    ]
+  },
+}
+
+export default command

src/commands/onboard-github/index.ts

@@ -4,7 +4,7 @@ const onboardGithub: Command = {
   name: 'onboard-github',
   aliases: ['onboarding-github', 'onboardgithub', 'onboardinggithub'],
   description:
-    'Interactive setup for GitHub Models: device login or PAT, saved to secure storage',
+    'Interactive setup for GitHub Copilot: OAuth device login stored in secure storage',
   type: 'local-jsx',
   load: () => import('./onboard-github.js'),
 }

src/commands/onboard-github/onboard-github.tsx

@@ -2,9 +2,9 @@ import * as React from 'react'
 import { useCallback, useState } from 'react'
 import { Select } from '../../components/CustomSelect/select.js'
 import { Spinner } from '../../components/Spinner.js'
-import TextInput from '../../components/TextInput.js'
 import { Box, Text } from '../../ink.js'
 import {
+  exchangeForCopilotToken,
   openVerificationUri,
   pollAccessToken,
   requestDeviceCode,
@@ -15,7 +15,7 @@ import {
   readGithubModelsToken,
   saveGithubModelsToken,
 } from '../../utils/githubModelsCredentials.js'
-import { updateSettingsForSource } from '../../utils/settings/settings.js'
+import { getSettingsForSource, updateSettingsForSource } from '../../utils/settings/settings.js'
 
 const DEFAULT_MODEL = 'github:copilot'
 const FORCE_RELOGIN_ARGS = new Set([
@@ -27,11 +27,25 @@ const FORCE_RELOGIN_ARGS = new Set([
   '--reauth',
 ])
 
-type Step =
+type Step = 'menu' | 'device-busy' | 'error'
-  | 'menu'
+
-  | 'device-busy'
+const PROVIDER_SPECIFIC_KEYS = new Set([
-  | 'pat'
+  'CLAUDE_CODE_USE_OPENAI',
-  | 'error'
+  'CLAUDE_CODE_USE_GEMINI',
+  'CLAUDE_CODE_USE_BEDROCK',
+  'CLAUDE_CODE_USE_VERTEX',
+  'CLAUDE_CODE_USE_FOUNDRY',
+  'OPENAI_BASE_URL',
+  'OPENAI_API_BASE',
+  'OPENAI_API_KEY',
+  'OPENAI_MODEL',
+  'GEMINI_API_KEY',
+  'GOOGLE_API_KEY',
+  'GEMINI_BASE_URL',
+  'GEMINI_MODEL',
+  'GEMINI_ACCESS_TOKEN',
+  'GEMINI_AUTH_MODE',
+])
 
 export function shouldForceGithubRelogin(args?: string): boolean {
   const normalized = (args ?? '').trim().toLowerCase()
@@ -41,15 +55,29 @@ export function shouldForceGithubRelogin(args?: string): boolean {
   return normalized.split(/\s+/).some(arg => FORCE_RELOGIN_ARGS.has(arg))
 }
 
+const GITHUB_PAT_PREFIXES = ['ghp_', 'gho_','ghs_', 'ghr_', 'github_pat_']
+
+function isGithubPat(token: string): boolean {
+  return GITHUB_PAT_PREFIXES.some(prefix => token.startsWith(prefix))
+}
+
 export function hasExistingGithubModelsLoginToken(
   env: NodeJS.ProcessEnv = process.env,
   storedToken?: string,
 ): boolean {
   const envToken = env.GITHUB_TOKEN?.trim() || env.GH_TOKEN?.trim()
   if (envToken) {
+    // PATs are no longer supported - require OAuth re-auth
+    if (isGithubPat(envToken)) {
+      return false
+    }
     return true
   }
   const persisted = (storedToken ?? readGithubModelsToken())?.trim()
+  // PATs are no longer supported - require OAuth re-auth
+  if (persisted && isGithubPat(persisted)) {
+    return false
+  }
   return Boolean(persisted)
 }
 
@@ -97,8 +125,21 @@ export function applyGithubOnboardingProcessEnv(
 }
 
 function mergeUserSettingsEnv(model: string): { ok: boolean; detail?: string } {
+  const currentSettings = getSettingsForSource('userSettings')
+  const currentEnv = currentSettings?.env ?? {}
+
+  const newEnv: Record<string, string> = {}
+  for (const [key, value] of Object.entries(currentEnv)) {
+    if (!PROVIDER_SPECIFIC_KEYS.has(key)) {
+      newEnv[key] = value
+    }
+  }
+
+  newEnv.CLAUDE_CODE_USE_GITHUB = '1'
+  newEnv.OPENAI_MODEL = model
+
   const { error } = updateSettingsForSource('userSettings', {
-    env: buildGithubOnboardingSettingsEnv(model) as any,
+    env: newEnv,
   })
   if (error) {
     return { ok: false, detail: error.message }
@@ -143,12 +184,14 @@ function OnboardGithub(props: {
     user_code: string
     verification_uri: string
   } | null>(null)
-  const [patDraft, setPatDraft] = useState('')
-  const [cursorOffset, setCursorOffset] = useState(0)
 
   const finalize = useCallback(
-    async (token: string, model: string = DEFAULT_MODEL) => {
+    async (
-      const saved = saveGithubModelsToken(token)
+      token: string,
+      model: string = DEFAULT_MODEL,
+      oauthToken?: string,
+    ) => {
+      const saved = saveGithubModelsToken(token, oauthToken)
       if (!saved.success) {
         setErrorMsg(saved.warning ?? 'Could not save token to secure storage.')
         setStep('error')
@@ -165,8 +208,18 @@ function OnboardGithub(props: {
         setStep('error')
         return
       }
+      // Clear stale provider-specific env vars from the current session
+      // so resolveProviderRequest() doesn't pick up a previous provider's
+      // base URL or key after onboarding completes.
+      for (const key of PROVIDER_SPECIFIC_KEYS) {
+        delete process.env[key]
+      }
+      process.env.CLAUDE_CODE_USE_GITHUB = '1'
+      process.env.OPENAI_MODEL = model.trim() || DEFAULT_MODEL
+      hydrateGithubModelsTokenFromSecureStorage()
+      onChangeAPIKey()
       onDone(
-        'GitHub Models onboard complete. Token stored in secure storage; user settings updated. Restart if the model does not switch.',
+        'GitHub Copilot onboard complete. Copilot token and OAuth token stored in secure storage (Windows/Linux: ~/.claude/.credentials.json, macOS: Keychain fallback to ~/.claude/.credentials.json); user settings updated. Restart if the model does not switch.',
         { display: 'user' },
       )
     },
@@ -184,11 +237,12 @@ function OnboardGithub(props: {
         verification_uri: device.verification_uri,
       })
       await openVerificationUri(device.verification_uri)
-      const token = await pollAccessToken(device.device_code, {
+      const oauthToken = await pollAccessToken(device.device_code, {
         initialInterval: device.interval,
         timeoutSeconds: device.expires_in,
       })
-      await finalize(token, DEFAULT_MODEL)
+      const copilotToken = await exchangeForCopilotToken(oauthToken)
+      await finalize(copilotToken.token, DEFAULT_MODEL, oauthToken)
     } catch (e) {
       setErrorMsg(e instanceof Error ? e.message : String(e))
       setStep('error')
@@ -227,7 +281,7 @@ function OnboardGithub(props: {
   if (step === 'device-busy') {
     return (
       <Box flexDirection="column" gap={1}>
-        <Text>GitHub device login</Text>
+        <Text>GitHub Copilot sign-in</Text>
         {deviceHint ? (
           <>
             <Text>
@@ -246,43 +300,11 @@ function OnboardGithub(props: {
     )
   }
 
-  if (step === 'pat') {
-    return (
-      <Box flexDirection="column" gap={1}>
-        <Text>Paste a GitHub personal access token with access to GitHub Models.</Text>
-        <Text dimColor>Input is masked. Enter to submit; Esc to go back.</Text>
-        <TextInput
-          value={patDraft}
-          mask="*"
-          onChange={setPatDraft}
-          onSubmit={async (value: string) => {
-            const t = value.trim()
-            if (!t) {
-              return
-            }
-            await finalize(t, DEFAULT_MODEL)
-          }}
-          onExit={() => {
-            setStep('menu')
-            setPatDraft('')
-          }}
-          columns={80}
-          cursorOffset={cursorOffset}
-          onChangeCursorOffset={setCursorOffset}
-        />
-      </Box>
-    )
-  }
-
   const menuOptions = [
     {
-      label: 'Sign in with browser (device code)',
+      label: 'Sign in with browser',
       value: 'device' as const,
     },
-    {
-      label: 'Paste personal access token',
-      value: 'pat' as const,
-    },
     {
       label: 'Cancel',
       value: 'cancel' as const,
@@ -291,7 +313,7 @@ function OnboardGithub(props: {
 
   return (
     <Box flexDirection="column" gap={1}>
-      <Text bold>GitHub Models setup</Text>
+      <Text bold>GitHub Copilot setup</Text>
       <Text dimColor>
         Stores your token in the OS credential store (macOS Keychain when available)
         and enables CLAUDE_CODE_USE_GITHUB in your user settings - no export
@@ -304,10 +326,6 @@ function OnboardGithub(props: {
             onDone('GitHub onboard cancelled', { display: 'system' })
             return
           }
-          if (v === 'pat') {
-            setStep('pat')
-            return
-          }
           void runDeviceFlow()
         }}
       />

src/commands/provider/provider.test.tsx

@@ -52,7 +52,11 @@ async function renderFinalFrame(node: React.ReactNode): Promise<string> {
     patchConsole: false,
   })
 
-  await instance.waitUntilExit()
+  // Timeout guard: if render throws before exit effect fires, don't hang
+  await Promise.race([
+    instance.waitUntilExit(),
+    new Promise<void>(resolve => setTimeout(resolve, 3000)),
+  ])
   return stripAnsi(extractLastFrame(getOutput()))
 }
 

src/components/CustomSelect/use-multi-select-state.ts

@@ -1,5 +1,4 @@
 import { useCallback, useState } from 'react'
-import { isDeepStrictEqual } from 'util'
 import { useRegisterOverlay } from '../../context/overlayContext.js'
 import type { InputEvent } from '../../ink/events/input-event.js'
 // eslint-disable-next-line custom-rules/prefer-use-keybindings -- raw space/arrow multiselect input
@@ -9,6 +8,7 @@ import {
   normalizeFullWidthSpace,
 } from '../../utils/stringUtils.js'
 import type { OptionWithDescription } from './select.js'
+import { optionsNavigateEqual } from './use-select-navigation.js'
 import { useSelectNavigation } from './use-select-navigation.js'
 
 export type UseMultiSelectStateProps<T> = {
@@ -174,7 +174,7 @@ export function useMultiSelectState<T>({
   // and the deleted ui/useMultiSelectState.ts — without this, MCPServerDesktopImportDialog
   // keeps colliding servers checked after getAllMcpConfigs() resolves.
   const [lastOptions, setLastOptions] = useState(options)
-  if (options !== lastOptions && !isDeepStrictEqual(options, lastOptions)) {
+  if (options !== lastOptions && !optionsNavigateEqual(options, lastOptions)) {
     setSelectedValues(defaultValue)
     setLastOptions(options)
   }

src/components/CustomSelect/use-select-navigation.ts

@@ -6,10 +6,34 @@ import {
   useRef,
   useState,
 } from 'react'
-import { isDeepStrictEqual } from 'util'
 import OptionMap from './option-map.js'
 import type { OptionWithDescription } from './select.js'
 
+/**
+ * Compare two option arrays for structural equality on properties that
+ * affect navigation behavior. ReactNode `label` and function `onChange`
+ * are intentionally excluded — they are identity-unstable (new reference
+ * each render) but don't change navigation semantics.
+ */
+export function optionsNavigateEqual<T>(
+  a: OptionWithDescription<T>[],
+  b: OptionWithDescription<T>[],
+): boolean {
+  if (a.length !== b.length) return false
+  for (let i = 0; i < a.length; i++) {
+    const ao = a[i]!
+    const bo = b[i]!
+    if (
+      ao.value !== bo.value ||
+      ao.disabled !== bo.disabled ||
+      ao.type !== bo.type
+    ) {
+      return false
+    }
+  }
+  return true
+}
+
 type State<T> = {
   /**
    * Map where key is option's value and value is option's index.
@@ -524,7 +548,7 @@ export function useSelectNavigation<T>({
 
   const [lastOptions, setLastOptions] = useState(options)
 
-  if (options !== lastOptions && !isDeepStrictEqual(options, lastOptions)) {
+  if (options !== lastOptions && !optionsNavigateEqual(options, lastOptions)) {
     dispatch({
       type: 'reset',
       state: createDefaultState({

src/components/ProviderManager.test.tsx

@@ -0,0 +1,305 @@
+import { PassThrough } from 'node:stream'
+
+import { afterEach, expect, mock, test } from 'bun:test'
+import React from 'react'
+import stripAnsi from 'strip-ansi'
+
+import { createRoot } from '../ink.js'
+import { AppStateProvider } from '../state/AppState.js'
+
+const SYNC_START = '\x1B[?2026h'
+const SYNC_END = '\x1B[?2026l'
+
+const ORIGINAL_ENV = {
+  CLAUDE_CODE_USE_GITHUB: process.env.CLAUDE_CODE_USE_GITHUB,
+  GITHUB_TOKEN: process.env.GITHUB_TOKEN,
+  GH_TOKEN: process.env.GH_TOKEN,
+}
+
+function extractLastFrame(output: string): string {
+  let lastFrame: string | null = null
+  let cursor = 0
+
+  while (cursor < output.length) {
+    const start = output.indexOf(SYNC_START, cursor)
+    if (start === -1) {
+      break
+    }
+
+    const contentStart = start + SYNC_START.length
+    const end = output.indexOf(SYNC_END, contentStart)
+    if (end === -1) {
+      break
+    }
+
+    const frame = output.slice(contentStart, end)
+    if (frame.trim().length > 0) {
+      lastFrame = frame
+    }
+    cursor = end + SYNC_END.length
+  }
+
+  return lastFrame ?? output
+}
+
+function createTestStreams(): {
+  stdout: PassThrough
+  stdin: PassThrough & {
+    isTTY: boolean
+    setRawMode: (mode: boolean) => void
+    ref: () => void
+    unref: () => void
+  }
+  getOutput: () => string
+} {
+  let output = ''
+  const stdout = new PassThrough()
+  const stdin = new PassThrough() as PassThrough & {
+    isTTY: boolean
+    setRawMode: (mode: boolean) => void
+    ref: () => void
+    unref: () => void
+  }
+
+  stdin.isTTY = true
+  stdin.setRawMode = () => {}
+  stdin.ref = () => {}
+  stdin.unref = () => {}
+  ;(stdout as unknown as { columns: number }).columns = 120
+  stdout.on('data', chunk => {
+    output += chunk.toString()
+  })
+
+  return {
+    stdout,
+    stdin,
+    getOutput: () => output,
+  }
+}
+
+async function waitForCondition(
+  predicate: () => boolean,
+  options?: { timeoutMs?: number; intervalMs?: number },
+): Promise<void> {
+  const timeoutMs = options?.timeoutMs ?? 2000
+  const intervalMs = options?.intervalMs ?? 10
+  const startedAt = Date.now()
+
+  while (Date.now() - startedAt < timeoutMs) {
+    if (predicate()) {
+      return
+    }
+    await Bun.sleep(intervalMs)
+  }
+
+  throw new Error('Timed out waiting for ProviderManager test condition')
+}
+
+function createDeferred<T>(): {
+  promise: Promise<T>
+  resolve: (value: T) => void
+} {
+  let resolve!: (value: T) => void
+  const promise = new Promise<T>(r => {
+    resolve = r
+  })
+  return { promise, resolve }
+}
+
+function mockProviderProfilesModule(): void {
+  mock.module('../utils/providerProfiles.js', () => ({
+    addProviderProfile: () => null,
+    applyActiveProviderProfileFromConfig: () => {},
+    deleteProviderProfile: () => ({ removed: false, activeProfileId: null }),
+    getActiveProviderProfile: () => null,
+    getProviderPresetDefaults: () => ({
+      provider: 'openai',
+      name: 'Mock provider',
+      baseUrl: 'http://localhost:11434/v1',
+      model: 'mock-model',
+      apiKey: '',
+    }),
+    getProviderProfiles: () => [],
+    setActiveProviderProfile: () => null,
+    updateProviderProfile: () => null,
+  }))
+}
+
+function mockProviderManagerDependencies(
+  syncRead: () => string | undefined,
+  asyncRead: () => Promise<string | undefined>,
+): void {
+  mockProviderProfilesModule()
+
+  mock.module('../utils/githubModelsCredentials.js', () => ({
+    clearGithubModelsToken: () => ({ success: true }),
+    GITHUB_MODELS_HYDRATED_ENV_MARKER: 'CLAUDE_CODE_GITHUB_TOKEN_HYDRATED',
+    hydrateGithubModelsTokenFromSecureStorage: () => {},
+    readGithubModelsToken: syncRead,
+    readGithubModelsTokenAsync: asyncRead,
+  }))
+
+  mock.module('../utils/settings/settings.js', () => ({
+    updateSettingsForSource: () => ({ error: null }),
+  }))
+}
+
+async function waitForFrameOutput(
+  getOutput: () => string,
+  predicate: (output: string) => boolean,
+  timeoutMs = 2500,
+): Promise<string> {
+  let output = ''
+
+  await waitForCondition(() => {
+    output = stripAnsi(extractLastFrame(getOutput()))
+    return predicate(output)
+  }, { timeoutMs })
+
+  return output
+}
+
+async function mountProviderManager(
+  ProviderManager: React.ComponentType<{
+    mode: 'first-run' | 'manage'
+    onDone: () => void
+  }>,
+): Promise<{
+  getOutput: () => string
+  dispose: () => Promise<void>
+}> {
+  const { stdout, stdin, getOutput } = createTestStreams()
+  const root = await createRoot({
+    stdout: stdout as unknown as NodeJS.WriteStream,
+    stdin: stdin as unknown as NodeJS.ReadStream,
+    patchConsole: false,
+  })
+
+  root.render(
+    <AppStateProvider>
+      <ProviderManager
+        mode="manage"
+        onDone={() => {}}
+      />
+    </AppStateProvider>,
+  )
+
+  return {
+    getOutput,
+    dispose: async () => {
+      root.unmount()
+      stdin.end()
+      stdout.end()
+      await Bun.sleep(0)
+    },
+  }
+}
+
+async function renderProviderManagerFrame(
+  ProviderManager: React.ComponentType<{
+    mode: 'first-run' | 'manage'
+    onDone: () => void
+  }>,
+  options?: {
+    waitForOutput?: (output: string) => boolean
+    timeoutMs?: number
+  },
+): Promise<string> {
+  const mounted = await mountProviderManager(ProviderManager)
+  const output = await waitForFrameOutput(
+    mounted.getOutput,
+    frame => {
+      if (!options?.waitForOutput) {
+        return frame.includes('Provider manager')
+      }
+      return options.waitForOutput(frame)
+    },
+    options?.timeoutMs ?? 2500,
+  )
+
+  await mounted.dispose()
+  return output
+}
+
+afterEach(() => {
+  mock.restore()
+
+  for (const [key, value] of Object.entries(ORIGINAL_ENV)) {
+    if (value === undefined) {
+      delete process.env[key as keyof typeof ORIGINAL_ENV]
+    } else {
+      process.env[key as keyof typeof ORIGINAL_ENV] = value
+    }
+  }
+})
+
+test('ProviderManager resolves GitHub virtual provider from async storage without sync reads in render flow', async () => {
+  delete process.env.CLAUDE_CODE_USE_GITHUB
+  delete process.env.GITHUB_TOKEN
+  delete process.env.GH_TOKEN
+
+  const syncRead = mock(() => {
+    throw new Error('sync credential read should not run in ProviderManager render flow')
+  })
+  const asyncRead = mock(async () => 'stored-token')
+
+  mockProviderManagerDependencies(syncRead, asyncRead)
+
+  const nonce = `${Date.now()}-${Math.random()}`
+  const { ProviderManager } = await import(`./ProviderManager.js?ts=${nonce}`)
+  const output = await renderProviderManagerFrame(ProviderManager, {
+    waitForOutput: frame =>
+      frame.includes('Provider manager') &&
+      frame.includes('GitHub Models') &&
+      frame.includes('token stored'),
+  })
+
+  expect(output).toContain('Provider manager')
+  expect(output).toContain('GitHub Models')
+  expect(output).toContain('token stored')
+  expect(output).not.toContain('No provider profiles configured yet.')
+
+  expect(syncRead).not.toHaveBeenCalled()
+  expect(asyncRead).toHaveBeenCalled()
+})
+
+test('ProviderManager avoids first-frame false negative while stored-token lookup is pending', async () => {
+  delete process.env.CLAUDE_CODE_USE_GITHUB
+  delete process.env.GITHUB_TOKEN
+  delete process.env.GH_TOKEN
+
+  const syncRead = mock(() => {
+    throw new Error('sync credential read should not run in ProviderManager render flow')
+  })
+  const deferredStoredToken = createDeferred<string | undefined>()
+  const asyncRead = mock(async () => deferredStoredToken.promise)
+
+  mockProviderManagerDependencies(syncRead, asyncRead)
+
+  const nonce = `${Date.now()}-${Math.random()}`
+  const { ProviderManager } = await import(`./ProviderManager.js?ts=${nonce}`)
+  const mounted = await mountProviderManager(ProviderManager)
+
+  const firstFrame = await waitForFrameOutput(
+    mounted.getOutput,
+    frame => frame.includes('Provider manager'),
+  )
+
+  expect(firstFrame).toContain('Checking GitHub Models credentials...')
+  expect(firstFrame).not.toContain('No provider profiles configured yet.')
+
+  deferredStoredToken.resolve('stored-token')
+
+  const resolvedFrame = await waitForFrameOutput(
+    mounted.getOutput,
+    frame => frame.includes('GitHub Models') && frame.includes('token stored'),
+  )
+
+  expect(resolvedFrame).toContain('GitHub Models')
+  expect(resolvedFrame).toContain('token stored')
+
+  await mounted.dispose()
+
+  expect(syncRead).not.toHaveBeenCalled()
+  expect(asyncRead).toHaveBeenCalled()
+})

src/components/ProviderManager.tsx

@@ -20,6 +20,7 @@ import {
   GITHUB_MODELS_HYDRATED_ENV_MARKER,
   hydrateGithubModelsTokenFromSecureStorage,
   readGithubModelsToken,
+  readGithubModelsTokenAsync,
 } from '../utils/githubModelsCredentials.js'
 import { isEnvTruthy } from '../utils/envUtils.js'
 import { updateSettingsForSource } from '../utils/settings/settings.js'
@@ -118,25 +119,38 @@ function profileSummary(profile: ProviderProfile, isActive: boolean): string {
   return `${providerKind} · ${profile.baseUrl} · ${profile.model} · ${keyInfo}${activeSuffix}`
 }
 
-function getGithubCredentialSource(
+function getGithubCredentialSourceFromEnv(
   processEnv: NodeJS.ProcessEnv = process.env,
 ): GithubCredentialSource {
-  if (readGithubModelsToken()?.trim()) {
-    return 'stored'
-  }
   if (processEnv.GITHUB_TOKEN?.trim() || processEnv.GH_TOKEN?.trim()) {
     return 'env'
   }
   return 'none'
 }
 
+async function resolveGithubCredentialSource(
+  processEnv: NodeJS.ProcessEnv = process.env,
+): Promise<GithubCredentialSource> {
+  const envSource = getGithubCredentialSourceFromEnv(processEnv)
+  if (envSource !== 'none') {
+    return envSource
+  }
+
+  if (await readGithubModelsTokenAsync()) {
+    return 'stored'
+  }
+
+  return 'none'
+}
+
 function isGithubProviderAvailable(
+  credentialSource: GithubCredentialSource,
   processEnv: NodeJS.ProcessEnv = process.env,
 ): boolean {
   if (isEnvTruthy(processEnv.CLAUDE_CODE_USE_GITHUB)) {
     return true
   }
-  return getGithubCredentialSource(processEnv) !== 'none'
+  return credentialSource !== 'none'
 }
 
 function getGithubProviderModel(
@@ -164,19 +178,24 @@ function getGithubProviderSummary(
 }
 
 export function ProviderManager({ mode, onDone }: Props): React.ReactNode {
+  const initialGithubCredentialSource = getGithubCredentialSourceFromEnv()
+  const initialIsGithubActive = isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB)
+  const initialHasGithubCredential = initialGithubCredentialSource !== 'none'
+
   const [profiles, setProfiles] = React.useState(() => getProviderProfiles())
   const [activeProfileId, setActiveProfileId] = React.useState(
     () => getActiveProviderProfile()?.id,
   )
-  const [githubProviderAvailable, setGithubProviderAvailable] = React.useState(() =>
+  const [githubProviderAvailable, setGithubProviderAvailable] = React.useState(
-    isGithubProviderAvailable(),
+    () => isGithubProviderAvailable(initialGithubCredentialSource),
   )
   const [githubCredentialSource, setGithubCredentialSource] = React.useState<GithubCredentialSource>(
-    () => getGithubCredentialSource(),
+    () => initialGithubCredentialSource,
-  )
-  const [isGithubActive, setIsGithubActive] = React.useState(() =>
-    isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB),
   )
+  const [isGithubActive, setIsGithubActive] = React.useState(() => initialIsGithubActive)
+  const [isGithubCredentialSourceResolved, setIsGithubCredentialSourceResolved] =
+    React.useState(() => initialHasGithubCredential || initialIsGithubActive)
+  const githubRefreshEpochRef = React.useRef(0)
   const [screen, setScreen] = React.useState<Screen>(
     mode === 'first-run' ? 'select-preset' : 'menu',
   )
@@ -196,13 +215,48 @@ export function ProviderManager({ mode, onDone }: Props): React.ReactNode {
   const currentStepKey = currentStep.key
   const currentValue = draft[currentStepKey]
 
+  const refreshGithubProviderState = React.useCallback((): void => {
+    const envCredentialSource = getGithubCredentialSourceFromEnv()
+    const githubActive = isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB)
+    const canResolveFromEnv = githubActive || envCredentialSource !== 'none'
+
+    if (canResolveFromEnv) {
+      githubRefreshEpochRef.current += 1
+      setGithubCredentialSource(envCredentialSource)
+      setGithubProviderAvailable(isGithubProviderAvailable(envCredentialSource))
+      setIsGithubActive(githubActive)
+      setIsGithubCredentialSourceResolved(true)
+      return
+    }
+
+    setIsGithubCredentialSourceResolved(false)
+    const refreshEpoch = ++githubRefreshEpochRef.current
+    void (async () => {
+      const credentialSource = await resolveGithubCredentialSource()
+      if (refreshEpoch !== githubRefreshEpochRef.current) {
+        return
+      }
+
+      setGithubCredentialSource(credentialSource)
+      setGithubProviderAvailable(isGithubProviderAvailable(credentialSource))
+      setIsGithubActive(isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB))
+      setIsGithubCredentialSourceResolved(true)
+    })()
+  }, [])
+
+  React.useEffect(() => {
+    refreshGithubProviderState()
+
+    return () => {
+      githubRefreshEpochRef.current += 1
+    }
+  }, [refreshGithubProviderState])
+
   function refreshProfiles(): void {
     const nextProfiles = getProviderProfiles()
     setProfiles(nextProfiles)
     setActiveProfileId(getActiveProviderProfile()?.id)
-    setGithubProviderAvailable(isGithubProviderAvailable())
+    refreshGithubProviderState()
-    setGithubCredentialSource(getGithubCredentialSource())
-    setIsGithubActive(isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB))
   }
 
   function clearStartupProviderOverrideFromUserSettings(): string | null {
@@ -640,7 +694,11 @@ export function ProviderManager({ mode, onDone }: Props): React.ReactNode {
         {statusMessage && <Text>{statusMessage}</Text>}
         <Box flexDirection="column">
           {profiles.length === 0 && !githubProviderAvailable ? (
+            isGithubCredentialSourceResolved ? (
               <Text dimColor>No provider profiles configured yet.</Text>
+            ) : (
+              <Text dimColor>Checking GitHub Models credentials...</Text>
+            )
           ) : (
             <>
               {profiles.map(profile => (

src/components/StartupScreen.ts

@@ -95,8 +95,8 @@ function detectProvider(): { name: string; model: string; baseUrl: string; isLoc
   if (useGithub) {
     const model = process.env.OPENAI_MODEL || 'github:copilot'
     const baseUrl =
-      process.env.OPENAI_BASE_URL || 'https://models.github.ai/inference'
+      process.env.OPENAI_BASE_URL || 'https://api.githubcopilot.com'
-    return { name: 'GitHub Models', model, baseUrl, isLocal: false }
+    return { name: 'GitHub Copilot', model, baseUrl, isLocal: false }
   }
 
   if (useOpenAI) {

src/components/agents/generateAgent.ts

@@ -68,11 +68,11 @@ When a user describes what they want an agent to do, you will:
       assistant: "Now let me use the test-runner agent to run the tests"
     </example>
     - <example>
-      Context: User is creating an agent to respond to the word "hello" with a friendly jok.
+      Context: User is creating an agent for Claude Code product questions.
-      user: "Hello"
+      user: "How do I configure Claude Code hooks?"
-      assistant: "I'm going to use the ${AGENT_TOOL_NAME} tool to launch the greeting-responder agent to respond with a friendly joke"
+      assistant: "I'm going to use the ${AGENT_TOOL_NAME} tool to launch the claude-code-guide agent to answer the question"
       <commentary>
-      Since the user is greeting, use the greeting-responder agent to respond with a friendly joke. 
+      Since the user is asking how to use Claude Code, use the claude-code-guide agent.
       </commentary>
     </example>
   - If the user mentioned or implied that the agent should be used proactively, you should include examples of this.

src/entrypoints/cli.tsx

@@ -8,6 +8,34 @@ import {
   validateProviderEnvOrExit,
 } from '../utils/providerValidation.js'
 
+// OpenClaude: polyfill globalThis.File for Node < 20.
+// undici v7 references `File` at module evaluation time (webidl type
+// assertions). Node 18 lacks the global, causing a ReferenceError inside
+// the bundled __commonJS require chain which deadlocks the process when a
+// proxy is configured (configureGlobalAgents → require_undici).
+// eslint-disable-next-line custom-rules/no-top-level-side-effects
+if (typeof globalThis.File === 'undefined') {
+  try {
+    // Node 18.13+ exposes File in node:buffer but not as a global.
+    // eslint-disable-next-line @typescript-eslint/no-require-imports
+    const { File: NodeFile } = require('node:buffer')
+    // @ts-expect-error -- polyfilling missing global
+    globalThis.File = NodeFile
+  } catch {
+    // Absolute fallback: stub so `MakeTypeAssertion(File)` doesn't throw.
+    // @ts-expect-error -- minimal polyfill
+    globalThis.File = class File extends Blob {
+      name: string
+      lastModified: number
+      constructor(parts: BlobPart[], name: string, opts?: FilePropertyBag) {
+        super(parts, opts)
+        this.name = name
+        this.lastModified = opts?.lastModified ?? Date.now()
+      }
+    }
+  }
+}
+
 // OpenClaude: disable experimental API betas by default.
 // Tool search (defer_loading), global cache scope, and context management
 // require internal API support not available to external accounts → 500.
@@ -68,15 +96,16 @@ async function main(): Promise<void> {
     }
   }
 
+  // Enable configs first so we can read settings
   {
     const { enableConfigs } = await import('../utils/config.js')
     enableConfigs()
+  }
+
+  // Apply settings.env from user settings (includes GitHub provider settings from /onboard-github)
+  {
     const { applySafeConfigEnvironmentVariables } = await import('../utils/managedEnv.js')
     applySafeConfigEnvironmentVariables()
-    const { hydrateGeminiAccessTokenFromSecureStorage } = await import('../utils/geminiCredentials.js')
-    hydrateGeminiAccessTokenFromSecureStorage()
-    const { hydrateGithubModelsTokenFromSecureStorage } = await import('../utils/githubModelsCredentials.js')
-    hydrateGithubModelsTokenFromSecureStorage()
   }
 
   const startupEnv = await buildStartupEnvFromProfile({
@@ -93,6 +122,16 @@ async function main(): Promise<void> {
     }
   }
 
+  // Hydrate GitHub credentials after profile is applied so CLAUDE_CODE_USE_GITHUB from profile is available
+  {
+    const {
+      hydrateGithubModelsTokenFromSecureStorage,
+      refreshGithubModelsTokenIfNeeded,
+    } = await import('../utils/githubModelsCredentials.js')
+    await refreshGithubModelsTokenIfNeeded()
+    hydrateGithubModelsTokenFromSecureStorage()
+  }
+
   await validateProviderEnvOrExit()
 
   // Print the gradient startup screen before the Ink UI loads

src/grpc/server.ts

@@ -40,7 +40,7 @@ export class GrpcServer {
       grpc.ServerCredentials.createInsecure(),
       (error, boundPort) => {
         if (error) {
-          console.error('Failed to start gRPC server', error)
+          console.error('Failed to start gRPC server')
           return
         }
         console.log(`gRPC Server running at ${host}:${boundPort}`)
@@ -225,7 +225,7 @@ export class GrpcServer {
           call.end()
         }
       } catch (err: any) {
-        console.error("Error processing stream:", err)
+        console.error('Error processing stream')
         call.write({
           error: {
             message: err.message || "Internal server error",

src/ink/reconciler.ts

@@ -366,14 +366,12 @@ const reconciler = createReconciler<
   createTextInstance(
     text: string,
     _root: DOMElement,
-    hostContext: HostContext,
+    _hostContext: HostContext,
   ): TextNode {
-    if (!hostContext.isInsideText) {
+    // react-compiler memoization can reuse cached <Text> elements without
-      throw new Error(
+    // re-traversing getChildHostContext, so hostContext.isInsideText may be
-        `Text string "${text}" must be rendered inside <Text> component`,
+    // stale. Always create the text node — Ink will render it correctly
-      )
+    // regardless of the context tracking state.
-    }
-
     return createTextNode(text)
   },
   resetTextContent() {},

src/ink/termio/osc.test.ts

@@ -27,6 +27,21 @@ async function flushClipboardCopy(): Promise<void> {
   await new Promise(resolve => setTimeout(resolve, 0))
 }
 
+async function waitForExecCall(
+  command: string,
+  attempts = 20,
+): Promise<(typeof execFileNoThrowMock.mock.calls)[number] | undefined> {
+  for (let attempt = 0; attempt < attempts; attempt++) {
+    const call = execFileNoThrowMock.mock.calls.find(([cmd]) => cmd === command)
+    if (call) {
+      return call
+    }
+    await flushClipboardCopy()
+  }
+
+  return undefined
+}
+
 describe('Windows clipboard fallback', () => {
   beforeEach(() => {
     execFileNoThrowMock.mockClear()
@@ -62,9 +77,7 @@ describe('Windows clipboard fallback', () => {
     await setClipboard('Привет мир')
     await flushClipboardCopy()
 
-    const windowsCall = execFileNoThrowMock.mock.calls.find(
+    const windowsCall = await waitForExecCall('powershell')
-      ([cmd]) => cmd === 'powershell',
-    )
 
     expect(windowsCall?.[2]).toMatchObject({
       stdin: 'ignore',

src/screens/REPL.tsx

@@ -238,6 +238,7 @@ import { usePromptsFromClaudeInChrome } from 'src/hooks/usePromptsFromClaudeInCh
 import { getTipToShowOnSpinner, recordShownTip } from 'src/services/tips/tipScheduler.js';
 import type { Theme } from 'src/utils/theme.js';
 import { isPromptTypingSuppressionActive } from './replInputSuppression.js';
+import { shouldRunStartupChecks } from './replStartupGates.js';
 import { checkAndDisableBypassPermissionsIfNeeded, checkAndDisableAutoModeIfNeeded, useKickOffCheckAndDisableBypassPermissionsIfNeeded, useKickOffCheckAndDisableAutoModeIfNeeded } from 'src/utils/permissions/bypassPermissionsKillswitch.js';
 import { SandboxManager } from 'src/utils/sandbox/sandbox-adapter.js';
 import { SANDBOX_NETWORK_ACCESS_TOOL_NAME } from 'src/cli/structuredIO.js';
@@ -792,10 +793,8 @@ export function REPL({
   // accepts, and only then is the REPL component mounted and this effect runs.
   // This ensures that plugin installations from repository and user settings only
   // happen after explicit user consent to trust the current working directory.
-  useEffect(() => {
+  // Deferring startup checks is handled below (after promptTypingSuppressionActive
-    if (isRemoteSession) return;
+  // is declared) to avoid temporal dead zone issues.
-    void performStartupChecks(setAppState);
-  }, [setAppState, isRemoteSession]);
 
   // Allow Claude in Chrome MCP to send prompts through MCP notifications
   // and sync permission mode changes to the Chrome extension
@@ -1429,6 +1428,25 @@ export function REPL({
   const activeRemote = sshRemote.isRemoteMode ? sshRemote : directConnect.isRemoteMode ? directConnect : remoteSession;
   const [pastedContents, setPastedContents] = useState<Record<number, PastedContent>>({});
   const [submitCount, setSubmitCount] = useState(0);
+
+  // Defer startup checks until the user has submitted their first message.
+  // A timeout or grace period is insufficient (issue #363): if the user pauses
+  // before typing, startup checks can still fire and recommendation dialogs
+  // steal focus. Only the user's first submission guarantees the prompt was
+  // the first thing they interacted with.
+  const startupChecksStartedRef = React.useRef(false);
+  const hasHadFirstSubmission = (submitCount ?? 0) > 0;
+  useEffect(() => {
+    if (isRemoteSession) return;
+    if (startupChecksStartedRef.current) return;
+    if (!shouldRunStartupChecks({
+      isRemoteSession,
+      hasStarted: startupChecksStartedRef.current,
+      hasHadFirstSubmission,
+    })) return;
+    startupChecksStartedRef.current = true;
+    void performStartupChecks(setAppState);
+  }, [setAppState, isRemoteSession, hasHadFirstSubmission]);
   // Ref instead of state to avoid triggering React re-renders on every
   // streaming text_delta. The spinner reads this via its animation timer.
   const responseLengthRef = useRef(0);
@@ -2061,13 +2079,14 @@ export function REPL({
     if (allowDialogsWithAnimation && showRemoteCallout) return 'remote-callout';
 
     // LSP plugin recommendation (lowest priority - non-blocking suggestion)
-    if (allowDialogsWithAnimation && lspRecommendation) return 'lsp-recommendation';
+    // Suppress during startup window to prevent stealing focus from the prompt (issue #363)
+    if (allowDialogsWithAnimation && lspRecommendation && startupChecksStartedRef.current) return 'lsp-recommendation';
 
     // Plugin hint from CLI/SDK stderr (same priority band as LSP rec)
-    if (allowDialogsWithAnimation && hintRecommendation) return 'plugin-hint';
+    if (allowDialogsWithAnimation && hintRecommendation && startupChecksStartedRef.current) return 'plugin-hint';
 
     // Desktop app upsell (max 3 launches, lowest priority)
-    if (allowDialogsWithAnimation && showDesktopUpsellStartup) return 'desktop-upsell';
+    if (allowDialogsWithAnimation && showDesktopUpsellStartup && startupChecksStartedRef.current) return 'desktop-upsell';
     return undefined;
   }
   const focusedInputDialog = getFocusedInputDialog();

src/screens/replStartupGates.test.ts

@@ -0,0 +1,53 @@
+import { describe, expect, test } from 'bun:test'
+
+import { shouldRunStartupChecks } from './replStartupGates.js'
+
+describe('shouldRunStartupChecks', () => {
+  test('runs checks after first message submission', () => {
+    expect(shouldRunStartupChecks({
+      isRemoteSession: false,
+      hasStarted: false,
+      hasHadFirstSubmission: true,
+    })).toBe(true)
+  })
+
+  test('skips checks in remote sessions even after submission', () => {
+    expect(shouldRunStartupChecks({
+      isRemoteSession: true,
+      hasStarted: false,
+      hasHadFirstSubmission: true,
+    })).toBe(false)
+  })
+
+  test('skips checks if already started', () => {
+    expect(shouldRunStartupChecks({
+      isRemoteSession: false,
+      hasStarted: true,
+      hasHadFirstSubmission: true,
+    })).toBe(false)
+  })
+
+  test('does not run checks before first submission', () => {
+    expect(shouldRunStartupChecks({
+      isRemoteSession: false,
+      hasStarted: false,
+      hasHadFirstSubmission: false,
+    })).toBe(false)
+  })
+
+  test('does not run checks when idle before first submission', () => {
+    expect(shouldRunStartupChecks({
+      isRemoteSession: false,
+      hasStarted: false,
+      hasHadFirstSubmission: false,
+    })).toBe(false)
+  })
+
+  test('skips checks in remote session regardless of other conditions', () => {
+    expect(shouldRunStartupChecks({
+      isRemoteSession: true,
+      hasStarted: false,
+      hasHadFirstSubmission: false,
+    })).toBe(false)
+  })
+})

src/screens/replStartupGates.ts

@@ -0,0 +1,35 @@
+/**
+ * Startup gates for the REPL.
+ *
+ * Prevents startup plugin checks and recommendation dialogs from stealing
+ * focus before the user has interacted with the prompt.
+ *
+ * This addresses the root cause of issue #363: on mount, performStartupChecks
+ * triggers plugin loading, which populates trackedFiles, which triggers
+ * useLspPluginRecommendation to surface an LSP recommendation dialog. Since
+ * promptTypingSuppressionActive is false before the user has typed anything,
+ * getFocusedInputDialog() returns the dialog, unmounting PromptInput entirely.
+ *
+ * The fix gates startup checks on actual prompt interaction. A pure timeout
+ * or grace period is insufficient because pausing before typing would still
+ * allow dialogs to steal focus. Only the user's first submission guarantees
+ * the prompt is no longer in the vulnerable pre-interaction window.
+ */
+
+/**
+ * Determines whether startup checks should run.
+ *
+ * Startup checks are deferred until the user has submitted their first
+ * message. This guarantees the prompt was the first thing the user interacted
+ * with, so no recommendation dialog can steal focus before the first keystroke.
+ */
+export function shouldRunStartupChecks(options: {
+  isRemoteSession: boolean;
+  hasStarted: boolean;
+  hasHadFirstSubmission: boolean;
+}): boolean {
+  if (options.isRemoteSession) return false;
+  if (options.hasStarted) return false;
+  if (!options.hasHadFirstSubmission) return false;
+  return true;
+}

src/services/api/client.test.ts

@@ -18,6 +18,7 @@ const originalEnv = {
   GEMINI_API_KEY: process.env.GEMINI_API_KEY,
   GEMINI_MODEL: process.env.GEMINI_MODEL,
   GEMINI_BASE_URL: process.env.GEMINI_BASE_URL,
+  GEMINI_AUTH_MODE: process.env.GEMINI_AUTH_MODE,
   GOOGLE_API_KEY: process.env.GOOGLE_API_KEY,
   OPENAI_API_KEY: process.env.OPENAI_API_KEY,
   OPENAI_BASE_URL: process.env.OPENAI_BASE_URL,
@@ -32,6 +33,7 @@ beforeEach(() => {
   process.env.GEMINI_API_KEY = 'gemini-test-key'
   process.env.GEMINI_MODEL = 'gemini-2.0-flash'
   process.env.GEMINI_BASE_URL = 'https://gemini.example/v1beta/openai'
+  process.env.GEMINI_AUTH_MODE = 'api-key'
 
   delete process.env.GOOGLE_API_KEY
   delete process.env.OPENAI_API_KEY
@@ -47,6 +49,7 @@ afterEach(() => {
   process.env.GEMINI_API_KEY = originalEnv.GEMINI_API_KEY
   process.env.GEMINI_MODEL = originalEnv.GEMINI_MODEL
   process.env.GEMINI_BASE_URL = originalEnv.GEMINI_BASE_URL
+  process.env.GEMINI_AUTH_MODE = originalEnv.GEMINI_AUTH_MODE
   process.env.GOOGLE_API_KEY = originalEnv.GOOGLE_API_KEY
   process.env.OPENAI_API_KEY = originalEnv.OPENAI_API_KEY
   process.env.OPENAI_BASE_URL = originalEnv.OPENAI_BASE_URL

src/services/api/codexShim.test.ts

@@ -17,16 +17,23 @@ const tempDirs: string[] = []
 const originalEnv = {
   OPENAI_BASE_URL: process.env.OPENAI_BASE_URL,
   OPENAI_API_BASE: process.env.OPENAI_API_BASE,
+  CLAUDE_CODE_USE_GITHUB: process.env.CLAUDE_CODE_USE_GITHUB,
 }
 
 afterEach(() => {
+  if (originalEnv.OPENAI_BASE_URL === undefined) delete process.env.OPENAI_BASE_URL
+  else process.env.OPENAI_BASE_URL = originalEnv.OPENAI_BASE_URL
+
+  if (originalEnv.OPENAI_API_BASE === undefined) delete process.env.OPENAI_API_BASE
+  else process.env.OPENAI_API_BASE = originalEnv.OPENAI_API_BASE
+
+  if (originalEnv.CLAUDE_CODE_USE_GITHUB === undefined) delete process.env.CLAUDE_CODE_USE_GITHUB
+  else process.env.CLAUDE_CODE_USE_GITHUB = originalEnv.CLAUDE_CODE_USE_GITHUB
+
   while (tempDirs.length > 0) {
     const dir = tempDirs.pop()
     if (dir) rmSync(dir, { recursive: true, force: true })
   }
-
-  process.env.OPENAI_BASE_URL = originalEnv.OPENAI_BASE_URL
-  process.env.OPENAI_API_BASE = originalEnv.OPENAI_API_BASE
 })
 
 function createTempAuthJson(payload: Record<string, unknown>): string {
@@ -71,6 +78,7 @@ describe('Codex provider config', () => {
   test('resolves codexplan alias to Codex transport with reasoning', () => {
     delete process.env.OPENAI_BASE_URL
     delete process.env.OPENAI_API_BASE
+    delete process.env.CLAUDE_CODE_USE_GITHUB
 
     const resolved = resolveProviderRequest({ model: 'codexplan' })
     expect(resolved.transport).toBe('codex_responses')
@@ -201,6 +209,117 @@ describe('Codex request translation', () => {
     ])
   })
 
+  test('preserves Grep tool pattern field in Codex strict schemas', () => {
+    const tools = convertToolsToResponsesTools([
+      {
+        name: 'Grep',
+        description: 'Search file contents',
+        input_schema: {
+          type: 'object',
+          properties: {
+            pattern: { type: 'string', description: 'Search pattern' },
+            path: { type: 'string' },
+          },
+          required: ['pattern'],
+          additionalProperties: false,
+        },
+      },
+    ])
+
+    expect(tools).toEqual([
+      {
+        type: 'function',
+        name: 'Grep',
+        description: 'Search file contents',
+        parameters: {
+          type: 'object',
+          properties: {
+            pattern: { type: 'string', description: 'Search pattern' },
+            path: { type: 'string' },
+          },
+          required: ['pattern', 'path'],
+          additionalProperties: false,
+        },
+        strict: true,
+      },
+    ])
+  })
+
+  test('preserves Glob tool pattern field in Codex strict schemas', () => {
+    const tools = convertToolsToResponsesTools([
+      {
+        name: 'Glob',
+        description: 'Find files by pattern',
+        input_schema: {
+          type: 'object',
+          properties: {
+            pattern: { type: 'string', description: 'Glob pattern' },
+            path: { type: 'string' },
+          },
+          required: ['pattern'],
+          additionalProperties: false,
+        },
+      },
+    ])
+
+    expect(tools).toEqual([
+      {
+        type: 'function',
+        name: 'Glob',
+        description: 'Find files by pattern',
+        parameters: {
+          type: 'object',
+          properties: {
+            pattern: { type: 'string', description: 'Glob pattern' },
+            path: { type: 'string' },
+          },
+          required: ['pattern', 'path'],
+          additionalProperties: false,
+        },
+        strict: true,
+      },
+    ])
+  })
+
+  test('strips validator pattern keyword but keeps string field named pattern in Codex schemas', () => {
+    const tools = convertToolsToResponsesTools([
+      {
+        name: 'RegexProbe',
+        description: 'Probe regex schema handling',
+        input_schema: {
+          type: 'object',
+          properties: {
+            pattern: {
+              type: 'string',
+              pattern: '^[a-z]+$',
+            },
+          },
+          required: ['pattern'],
+          additionalProperties: false,
+        },
+      },
+    ])
+
+    expect(tools).toEqual([
+      {
+        type: 'function',
+        name: 'RegexProbe',
+        description: 'Probe regex schema handling',
+        parameters: {
+          type: 'object',
+          properties: {
+            pattern: {
+              type: 'string',
+            },
+          },
+          required: ['pattern'],
+          additionalProperties: false,
+        },
+        strict: true,
+      },
+    ])
+  })
+
   test('removes unsupported uri format from strict Responses schemas', () => {
     const tools = convertToolsToResponsesTools([
       {

src/services/api/openaiShim.ts

@@ -15,9 +15,9 @@
  *   OPENAI_MODEL=gpt-4o              — default model override
  *   CODEX_API_KEY / ~/.codex/auth.json — Codex auth for codexplan/codexspark
  *
- * GitHub Models (models.github.ai), OpenAI-compatible:
+ * GitHub Copilot API (api.githubcopilot.com), OpenAI-compatible:
  *   CLAUDE_CODE_USE_GITHUB=1         — enable GitHub inference (no need for USE_OPENAI)
- *   GITHUB_TOKEN or GH_TOKEN         — PAT with models access (mapped to Bearer auth)
+ *   GITHUB_TOKEN or GH_TOKEN         — Copilot API token (mapped to Bearer auth)
  *   OPENAI_MODEL                     — optional; use github:copilot or openai/gpt-4.1 style IDs
  */
 
@@ -29,7 +29,9 @@ import { hydrateGithubModelsTokenFromSecureStorage } from '../../utils/githubMod
 import {
   codexStreamToAnthropic,
   collectCodexCompletedResponse,
+  convertAnthropicMessagesToResponsesInput,
   convertCodexResponseToAnthropicMessage,
+  convertToolsToResponsesTools,
   performCodexRequest,
   type AnthropicStreamEvent,
   type AnthropicUsage,
@@ -39,9 +41,14 @@ import {
   isLocalProviderUrl,
   resolveCodexApiCredentials,
   resolveProviderRequest,
+  getGithubEndpointType,
 } from './providerConfig.js'
 import { sanitizeSchemaForOpenAICompat } from '../../utils/schemaSanitizer.js'
 import { redactSecretValueForDisplay } from '../../utils/providerProfile.js'
+import {
+  normalizeToolArguments,
+  hasToolFieldMapping,
+} from './toolArgumentNormalization.js'
 
 type SecretValueSource = Partial<{
   OPENAI_API_KEY: string
@@ -51,16 +58,33 @@ type SecretValueSource = Partial<{
   GEMINI_ACCESS_TOKEN: string
 }>
 
-const GITHUB_MODELS_DEFAULT_BASE = 'https://models.github.ai/inference'
+const GITHUB_COPILOT_BASE = 'https://api.githubcopilot.com'
-const GITHUB_API_VERSION = '2022-11-28'
 const GITHUB_429_MAX_RETRIES = 3
 const GITHUB_429_BASE_DELAY_SEC = 1
 const GITHUB_429_MAX_DELAY_SEC = 32
+const GEMINI_API_HOST = 'generativelanguage.googleapis.com'
+
+const COPILOT_HEADERS: Record<string, string> = {
+  'User-Agent': 'GitHubCopilotChat/0.26.7',
+  'Editor-Version': 'vscode/1.99.3',
+  'Editor-Plugin-Version': 'copilot-chat/0.26.7',
+  'Copilot-Integration-Id': 'vscode-chat',
+}
 
 function isGithubModelsMode(): boolean {
   return isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB)
 }
 
+function hasGeminiApiHost(baseUrl: string | undefined): boolean {
+  if (!baseUrl) return false
+
+  try {
+    return new URL(baseUrl).hostname.toLowerCase() === GEMINI_API_HOST
+  } catch {
+    return false
+  }
+}
+
 function formatRetryAfterHint(response: Response): string {
   const ra = response.headers.get('retry-after')
   return ra ? ` (Retry-After: ${ra})` : ''
@@ -180,10 +204,12 @@ function convertContentBlocks(
         // handled separately
         break
       case 'thinking':
-        // Append thinking as text with a marker for models that support reasoning
+      case 'redacted_thinking':
-        if (block.thinking) {
+        // Strip thinking blocks for OpenAI-compatible providers.
-          parts.push({ type: 'text', text: `<thinking>${block.thinking}</thinking>` })
+        // These are Anthropic-specific content types that 3P providers
-        }
+        // don't understand. Serializing them as <thinking> text corrupts
+        // multi-turn context: the model sees the tags as part of its
+        // previous reply and may mimic or misattribute them.
         break
       default:
         if (block.text) {
@@ -197,6 +223,13 @@ function convertContentBlocks(
   return parts
 }
 
+function isGeminiMode(): boolean {
+  return (
+    isEnvTruthy(process.env.CLAUDE_CODE_USE_GEMINI) ||
+    hasGeminiApiHost(process.env.OPENAI_BASE_URL)
+  )
+}
+
 function convertMessages(
   messages: Array<{ role: string; message?: { role?: string; content?: unknown }; content?: unknown }>,
   system: unknown,
@@ -248,6 +281,7 @@ function convertMessages(
       // Check for tool_use blocks
       if (Array.isArray(content)) {
         const toolUses = content.filter((b: { type?: string }) => b.type === 'tool_use')
+        const thinkingBlock = content.find((b: { type?: string }) => b.type === 'thinking')
         const textContent = content.filter(
           (b: { type?: string }) => b.type !== 'tool_use' && b.type !== 'thinking',
         )
@@ -267,7 +301,9 @@ function convertMessages(
               name?: string
               input?: unknown
               extra_content?: Record<string, unknown>
-            }) => ({
+              signature?: string
+            }, index) => {
+              const toolCall: NonNullable<OpenAIMessage['tool_calls']>[number] = {
                 id: tu.id ?? `call_${crypto.randomUUID().replace(/-/g, '')}`,
                 type: 'function' as const,
                 function: {
@@ -277,8 +313,34 @@ function convertMessages(
                       ? tu.input
                       : JSON.stringify(tu.input ?? {}),
                 },
-              ...(tu.extra_content ? { extra_content: tu.extra_content } : {}),
+              }
-            }),
+
+              // Preserve existing extra_content if present
+              if (tu.extra_content) {
+                toolCall.extra_content = { ...tu.extra_content }
+              }
+
+              // Handle Gemini thought_signature
+              if (isGeminiMode()) {
+                // If the model provided a signature in the tool_use block itself (e.g. from a previous Turn/Step)
+                // Use thinkingBlock.signature for ALL tool calls in the same assistant turn if available.
+                // The API requires the same signature on every replayed function call part in a parallel set.
+                const signature = tu.signature ?? (thinkingBlock as any)?.signature
+
+                // Merge into existing google-specific metadata if present
+                const existingGoogle = (toolCall.extra_content?.google as Record<string, unknown>) ?? {}
+
+                toolCall.extra_content = {
+                  ...toolCall.extra_content,
+                  google: {
+                    ...existingGoogle,
+                    thought_signature: signature ?? "skip_thought_signature_validator"
+                  }
+                }
+              }
+
+              return toolCall
+            },
           )
         }
 
@@ -359,11 +421,13 @@ function normalizeSchemaForOpenAI(
     record.properties = normalizedProps
 
     if (strict) {
-      // OpenAI strict mode requires every property to be listed in required[]
+      // Keep only the properties that were originally marked required in the schema.
-      const allKeys = Object.keys(normalizedProps)
+      // Adding every property to required[] (the previous behaviour) caused strict
-      record.required = Array.from(new Set([...existingRequired, ...allKeys]))
+      // OpenAI-compatible providers (Groq, Azure, etc.) to reject tool calls because
-      // OpenAI strict mode requires additionalProperties: false on all object
+      // the model correctly omits optional arguments — but the provider treats them
-      // schemas — override unconditionally to ensure nested objects comply.
+      // as missing required fields and returns a 400 / tool_use_failed error.
+      record.required = existingRequired.filter(k => k in normalizedProps)
+      // additionalProperties: false is still required by strict-mode providers.
       record.additionalProperties = false
     } else {
       // For Gemini: keep only existing required keys that are present in properties
@@ -397,7 +461,7 @@ function normalizeSchemaForOpenAI(
 function convertTools(
   tools: Array<{ name: string; description?: string; input_schema?: Record<string, unknown> }>,
 ): OpenAITool[] {
-  const isGemini = isEnvTruthy(process.env.CLAUDE_CODE_USE_GEMINI)
+  const isGemini = isGeminiMode()
 
   return tools
     .filter(t => t.name !== 'ToolSearchTool') // Not relevant for OpenAI
@@ -439,6 +503,7 @@ interface OpenAIStreamChunk {
     delta: {
       role?: string
       content?: string | null
+      reasoning_content?: string | null
       tool_calls?: Array<{
         index: number
         id?: string
@@ -476,6 +541,30 @@ function convertChunkUsage(
   }
 }
 
+const JSON_REPAIR_SUFFIXES = [
+  '}', '"}', ']}', '"]}', '}}', '"}}', ']}}', '"]}}', '"]}]}', '}]}'
+]
+
+function repairPossiblyTruncatedObjectJson(raw: string): string | null {
+  try {
+    const parsed = JSON.parse(raw)
+    return parsed && typeof parsed === 'object' && !Array.isArray(parsed)
+      ? raw
+      : null
+  } catch {
+    for (const combo of JSON_REPAIR_SUFFIXES) {
+      try {
+        const repaired = raw + combo
+        const parsed = JSON.parse(repaired)
+        if (parsed && typeof parsed === 'object' && !Array.isArray(parsed)) {
+          return repaired
+        }
+      } catch {}
+    }
+    return null
+  }
+}
+
 /**
  * Async generator that transforms an OpenAI SSE stream into
  * Anthropic-format BetaRawMessageStreamEvent objects.
@@ -486,8 +575,19 @@ async function* openaiStreamToAnthropic(
 ): AsyncGenerator<AnthropicStreamEvent> {
   const messageId = makeMessageId()
   let contentBlockIndex = 0
-  const activeToolCalls = new Map<number, { id: string; name: string; index: number; jsonBuffer: string }>()
+  const activeToolCalls = new Map<
+    number,
+    {
+      id: string
+      name: string
+      index: number
+      jsonBuffer: string
+      normalizeAtStop: boolean
+    }
+  >()
   let hasEmittedContentStart = false
+  let hasEmittedThinkingStart = false
+  let hasClosedThinking = false
   let lastStopReason: 'tool_use' | 'max_tokens' | 'end_turn' | null = null
   let hasEmittedFinalUsage = false
   let hasProcessedFinishReason = false
@@ -544,9 +644,34 @@ async function* openaiStreamToAnthropic(
       for (const choice of chunk.choices ?? []) {
         const delta = choice.delta
 
+        // Reasoning models (e.g. GLM-5, DeepSeek) may stream chain-of-thought
+        // in `reasoning_content` before the actual reply appears in `content`.
+        // Emit reasoning as a thinking block and content as a text block.
+        if (delta.reasoning_content != null && delta.reasoning_content !== '') {
+          if (!hasEmittedThinkingStart) {
+            yield {
+              type: 'content_block_start',
+              index: contentBlockIndex,
+              content_block: { type: 'thinking', thinking: '' },
+            }
+            hasEmittedThinkingStart = true
+          }
+          yield {
+            type: 'content_block_delta',
+            index: contentBlockIndex,
+            delta: { type: 'thinking_delta', thinking: delta.reasoning_content },
+          }
+        }
+
         // Text content — use != null to distinguish absent field from empty string,
         // some providers send "" as first delta to signal streaming start
-        if (delta.content != null) {
+        if (delta.content != null && delta.content !== '') {
+          // Close thinking block if transitioning from reasoning to content
+          if (hasEmittedThinkingStart && !hasClosedThinking) {
+            yield { type: 'content_block_stop', index: contentBlockIndex }
+            contentBlockIndex++
+            hasClosedThinking = true
+          }
           if (!hasEmittedContentStart) {
             yield {
               type: 'content_block_start',
@@ -566,7 +691,12 @@ async function* openaiStreamToAnthropic(
         if (delta.tool_calls) {
           for (const tc of delta.tool_calls) {
             if (tc.id && tc.function?.name) {
-              // New tool call starting
+              // New tool call starting — close any open thinking block first
+              if (hasEmittedThinkingStart && !hasClosedThinking) {
+                yield { type: 'content_block_stop', index: contentBlockIndex }
+                contentBlockIndex++
+                hasClosedThinking = true
+              }
               if (hasEmittedContentStart) {
                 yield {
                   type: 'content_block_stop',
@@ -577,11 +707,14 @@ async function* openaiStreamToAnthropic(
               }
 
               const toolBlockIndex = contentBlockIndex
+              const initialArguments = tc.function.arguments ?? ''
+              const normalizeAtStop = hasToolFieldMapping(tc.function.name)
               activeToolCalls.set(tc.index, {
                 id: tc.id,
                 name: tc.function.name,
                 index: toolBlockIndex,
-                jsonBuffer: tc.function.arguments ?? '',
+                jsonBuffer: initialArguments,
+                normalizeAtStop,
               })
 
               yield {
@@ -593,12 +726,19 @@ async function* openaiStreamToAnthropic(
                   name: tc.function.name,
                   input: {},
                   ...(tc.extra_content ? { extra_content: tc.extra_content } : {}),
+                  // Extract Gemini signature from extra_content
+                  ...((tc.extra_content?.google as any)?.thought_signature
+                    ? {
+                        signature: (tc.extra_content.google as any)
+                          .thought_signature,
+                      }
+                    : {}),
                 },
               }
               contentBlockIndex++
 
               // Emit any initial arguments
-              if (tc.function.arguments) {
+              if (tc.function.arguments && !normalizeAtStop) {
                 yield {
                   type: 'content_block_delta',
                   index: toolBlockIndex,
@@ -615,6 +755,11 @@ async function* openaiStreamToAnthropic(
                 if (tc.function.arguments) {
                   active.jsonBuffer += tc.function.arguments
                 }
+
+                if (active.normalizeAtStop) {
+                  continue
+                }
+
                 yield {
                   type: 'content_block_delta',
                   index: active.index,
@@ -633,6 +778,12 @@ async function* openaiStreamToAnthropic(
         if (choice.finish_reason && !hasProcessedFinishReason) {
           hasProcessedFinishReason = true
 
+          // Close any open thinking block that wasn't closed by content transition
+          if (hasEmittedThinkingStart && !hasClosedThinking) {
+            yield { type: 'content_block_stop', index: contentBlockIndex }
+            contentBlockIndex++
+            hasClosedThinking = true
+          }
           // Close any open content blocks
           if (hasEmittedContentStart) {
             yield {
@@ -642,16 +793,44 @@ async function* openaiStreamToAnthropic(
           }
           // Close active tool calls
           for (const [, tc] of activeToolCalls) {
+            if (tc.normalizeAtStop) {
+              let partialJson: string
+              if (choice.finish_reason === 'length') {
+                // Truncated by max tokens — preserve raw buffer to avoid
+                // turning an incomplete tool call into an executable command
+                partialJson = tc.jsonBuffer
+              } else {
+                const repairedStructuredJson = repairPossiblyTruncatedObjectJson(
+                  tc.jsonBuffer,
+                )
+                if (repairedStructuredJson) {
+                  partialJson = repairedStructuredJson
+                } else {
+                  partialJson = JSON.stringify(
+                    normalizeToolArguments(tc.name, tc.jsonBuffer),
+                  )
+                }
+              }
+
+              yield {
+                type: 'content_block_delta',
+                index: tc.index,
+                delta: {
+                  type: 'input_json_delta',
+                  partial_json: partialJson,
+                },
+              }
+              yield { type: 'content_block_stop', index: tc.index }
+              continue
+            }
+
             let suffixToAdd = ''
             if (tc.jsonBuffer) {
               try {
                 JSON.parse(tc.jsonBuffer)
               } catch {
                 const str = tc.jsonBuffer.trimEnd()
-                const combinations = [
+                for (const combo of JSON_REPAIR_SUFFIXES) {
-                  '}', '"}', ']}', '"]}', '}}', '"}}', ']}}', '"]}}', '"]}]}', '}]}'
-                ]
-                for (const combo of combinations) {
                   try {
                     JSON.parse(str + combo)
                     suffixToAdd = combo
@@ -776,8 +955,9 @@ class OpenAIShimMessages {
       httpResponse = response
 
       if (params.stream) {
+        const isResponsesStream = response.url?.includes('/responses')
         return new OpenAIShimStream(
-          request.transport === 'codex_responses'
+          (request.transport === 'codex_responses' || isResponsesStream)
             ? codexStreamToAnthropic(response, request.resolvedModel)
             : openaiStreamToAnthropic(response, request.resolvedModel),
         )
@@ -791,8 +971,38 @@ class OpenAIShimMessages {
         )
       }
 
+      const isResponsesNonStream = response.url?.includes('/responses')
+      if (isResponsesNonStream || (request.transport === 'chat_completions' && isGithubModelsMode())) {
+        const contentType = response.headers.get('content-type') ?? ''
+        if (contentType.includes('application/json')) {
+          const parsed = await response.json() as Record<string, unknown>
+          if (
+            parsed &&
+            typeof parsed === 'object' &&
+            ('output' in parsed || 'incomplete_details' in parsed)
+          ) {
+            return convertCodexResponseToAnthropicMessage(
+              parsed,
+              request.resolvedModel,
+            )
+          }
+          return self._convertNonStreamingResponse(parsed, request.resolvedModel)
+        }
+      }
+
+      const contentType = response.headers.get('content-type') ?? ''
+      if (contentType.includes('application/json')) {
         const data = await response.json()
         return self._convertNonStreamingResponse(data, request.resolvedModel)
+      }
+
+      const textBody = await response.text().catch(() => '')
+      throw APIError.generate(
+        response.status,
+        undefined,
+        `OpenAI API error ${response.status}: unexpected response: ${textBody.slice(0, 500)}`,
+        response.headers as unknown as Headers,
+      )
     })()
 
       ; (promise as unknown as Record<string, unknown>).withResponse =
@@ -814,7 +1024,36 @@ class OpenAIShimMessages {
     params: ShimCreateParams,
     options?: { signal?: AbortSignal; headers?: Record<string, string> },
   ): Promise<Response> {
-    if (request.transport === 'codex_responses') {
+    const githubEndpointType = getGithubEndpointType(request.baseUrl)
+    const isGithubMode = isGithubModelsMode()
+    const isGithubWithCodexTransport = isGithubMode && request.transport === 'codex_responses'
+    const isGithubCopilotEndpoint = isGithubMode && githubEndpointType === 'copilot'
+
+    if (isGithubWithCodexTransport) {
+      const apiKey = this.providerOverride?.apiKey ?? process.env.OPENAI_API_KEY ?? ''
+      if (!apiKey) {
+        throw new Error(
+          'GitHub Copilot auth is required. Run /onboard-github to sign in.',
+        )
+      }
+
+      return performCodexRequest({
+        request,
+        credentials: {
+          apiKey,
+          source: 'env',
+        },
+        params,
+        defaultHeaders: {
+          ...this.defaultHeaders,
+          ...(options?.headers ?? {}),
+          ...COPILOT_HEADERS,
+        },
+        signal: options?.signal,
+      })
+    }
+
+    if (request.transport === 'codex_responses' && !isGithubMode) {
       const credentials = resolveCodexApiCredentials()
       if (!credentials.apiKey) {
         const authHint = credentials.authPath
@@ -888,6 +1127,10 @@ class OpenAIShimMessages {
     }
 
     const isGithub = isGithubModelsMode()
+    const githubEndpointType = getGithubEndpointType(request.baseUrl)
+    const isGithubCopilot = isGithub && githubEndpointType === 'copilot'
+    const isGithubModels = isGithub && (githubEndpointType === 'models' || githubEndpointType === 'custom')
+
     if (isGithub && body.max_completion_tokens !== undefined) {
       body.max_tokens = body.max_completion_tokens
       delete body.max_completion_tokens
@@ -930,7 +1173,7 @@ class OpenAIShimMessages {
       ...(options?.headers ?? {}),
     }
 
-    const isGemini = isEnvTruthy(process.env.CLAUDE_CODE_USE_GEMINI)
+    const isGemini = isGeminiMode()
     const apiKey =
       this.providerOverride?.apiKey ?? process.env.OPENAI_API_KEY ?? ''
     // Detect Azure endpoints by hostname (not raw URL) to prevent bypass via
@@ -953,15 +1196,17 @@ class OpenAIShimMessages {
       const geminiCredential = await resolveGeminiCredential(process.env)
       if (geminiCredential.kind !== 'none') {
         headers.Authorization = `Bearer ${geminiCredential.credential}`
-        if (geminiCredential.projectId) {
+        if (geminiCredential.kind !== 'api-key' && 'projectId' in geminiCredential && geminiCredential.projectId) {
           headers['x-goog-user-project'] = geminiCredential.projectId
         }
       }
     }
 
-    if (isGithub) {
+    if (isGithubCopilot) {
-      headers.Accept = 'application/vnd.github.v3+json'
+      Object.assign(headers, COPILOT_HEADERS)
-      headers['X-GitHub-Api-Version'] = GITHUB_API_VERSION
+    } else if (isGithubModels) {
+      headers['Accept'] = 'application/vnd.github+json'
+      headers['X-GitHub-Api-Version'] = '2022-11-28'
     }
 
     // Build the chat completions URL
@@ -1013,9 +1258,82 @@ class OpenAIShimMessages {
         await sleepMs(delaySec * 1000)
         continue
       }
+      // Read body exactly once here — Response body is a stream that can only
+      // be consumed a single time.
       const errorBody = await response.text().catch(() => 'unknown error')
       const rateHint =
         isGithub && response.status === 429 ? formatRetryAfterHint(response) : ''
+
+      // If GitHub Copilot returns error about /chat/completions,
+      // try the /responses endpoint (needed for GPT-5+ models)
+      if (isGithub && response.status === 400) {
+        if (errorBody.includes('/chat/completions') || errorBody.includes('not accessible')) {
+          const responsesUrl = `${request.baseUrl}/responses`
+          const responsesBody: Record<string, unknown> = {
+            model: request.resolvedModel,
+            input: convertAnthropicMessagesToResponsesInput(
+              params.messages as Array<{
+                role?: string
+                message?: { role?: string; content?: unknown }
+                content?: unknown
+              }>,
+            ),
+            stream: params.stream ?? false,
+          }
+
+          if (!Array.isArray(responsesBody.input) || responsesBody.input.length === 0) {
+            responsesBody.input = [
+              {
+                type: 'message',
+                role: 'user',
+                content: [{ type: 'input_text', text: '' }],
+              },
+            ]
+          }
+
+          const systemText = convertSystemPrompt(params.system)
+          if (systemText) {
+            responsesBody.instructions = systemText
+          }
+
+          if (body.max_tokens !== undefined) {
+            responsesBody.max_output_tokens = body.max_tokens
+          }
+
+          if (params.tools && params.tools.length > 0) {
+            const convertedTools = convertToolsToResponsesTools(
+              params.tools as Array<{
+                name?: string
+                description?: string
+                input_schema?: Record<string, unknown>
+              }>,
+            )
+            if (convertedTools.length > 0) {
+              responsesBody.tools = convertedTools
+            }
+          }
+
+          const responsesResponse = await fetch(responsesUrl, {
+            method: 'POST',
+            headers,
+            body: JSON.stringify(responsesBody),
+            signal: options?.signal,
+          })
+          if (responsesResponse.ok) {
+            return responsesResponse
+          }
+          const responsesErrorBody = await responsesResponse.text().catch(() => 'unknown error')
+          let responsesErrorResponse: object | undefined
+          try { responsesErrorResponse = JSON.parse(responsesErrorBody) } catch { /* raw text */ }
+          throw APIError.generate(
+            responsesResponse.status,
+            responsesErrorResponse,
+            `OpenAI API error ${responsesResponse.status}: ${responsesErrorBody}`,
+            responsesResponse.headers,
+          )
+        }
+      }
+
       let errorResponse: object | undefined
       try { errorResponse = JSON.parse(errorBody) } catch { /* raw text */ }
       throw APIError.generate(
@@ -1043,6 +1361,7 @@ class OpenAIShimMessages {
             | string
             | null
             | Array<{ type?: string; text?: string }>
+          reasoning_content?: string | null
           tool_calls?: Array<{
             id: string
             function: { name: string; arguments: string }
@@ -1064,7 +1383,17 @@ class OpenAIShimMessages {
     const choice = data.choices?.[0]
     const content: Array<Record<string, unknown>> = []
 
-    const rawContent = choice?.message?.content
+    // Some reasoning models (e.g. GLM-5) put their reply in reasoning_content
+    // while content stays null — emit reasoning as a thinking block, then
+    // fall back to it for visible text if content is empty.
+    const reasoningText = choice?.message?.reasoning_content
+    if (typeof reasoningText === 'string' && reasoningText) {
+      content.push({ type: 'thinking', thinking: reasoningText })
+    }
+    const rawContent =
+      choice?.message?.content !== '' && choice?.message?.content != null
+        ? choice?.message?.content
+        : choice?.message?.reasoning_content
     if (typeof rawContent === 'string' && rawContent) {
       content.push({ type: 'text', text: rawContent })
     } else if (Array.isArray(rawContent) && rawContent.length > 0) {
@@ -1087,18 +1416,20 @@ class OpenAIShimMessages {
 
     if (choice?.message?.tool_calls) {
       for (const tc of choice.message.tool_calls) {
-        let input: unknown
+        const input = normalizeToolArguments(
-        try {
+          tc.function.name,
-          input = JSON.parse(tc.function.arguments)
+          tc.function.arguments,
-        } catch {
+        )
-          input = { raw: tc.function.arguments }
-        }
         content.push({
           type: 'tool_use',
           id: tc.id,
           name: tc.function.name,
           input,
           ...(tc.extra_content ? { extra_content: tc.extra_content } : {}),
+          // Extract Gemini signature from extra_content
+          ...((tc.extra_content?.google as any)?.thought_signature
+            ? { signature: (tc.extra_content.google as any).thought_signature }
+            : {}),
         })
       }
     }
@@ -1170,7 +1501,7 @@ export function createOpenAIShimClient(options: {
       process.env.OPENAI_MODEL = process.env.GEMINI_MODEL
     }
   } else if (isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB)) {
-    process.env.OPENAI_BASE_URL ??= GITHUB_MODELS_DEFAULT_BASE
+    process.env.OPENAI_BASE_URL ??= GITHUB_COPILOT_BASE
     process.env.OPENAI_API_KEY ??=
       process.env.GITHUB_TOKEN ?? process.env.GH_TOKEN ?? ''
   }

src/services/api/providerConfig.github.test.ts

@@ -23,6 +23,9 @@ test.each([
   ['github:gpt-4o', 'gpt-4o'],
   ['gpt-4o', 'gpt-4o'],
   ['github:copilot?reasoning=high', DEFAULT_GITHUB_MODELS_API_MODEL],
+  // normalizeGithubModelsApiModel preserves provider prefix for models.github.ai compatibility
+  ['github:openai/gpt-4.1', 'openai/gpt-4.1'],
+  ['openai/gpt-4.1', 'openai/gpt-4.1'],
 ] as const)('normalizeGithubModelsApiModel(%s) -> %s', (input, expected) => {
   expect(normalizeGithubModelsApiModel(input)).toBe(expected)
 })
@@ -34,6 +37,20 @@ test('resolveProviderRequest applies GitHub normalization when CLAUDE_CODE_USE_G
   expect(r.transport).toBe('chat_completions')
 })
 
+test('resolveProviderRequest routes GitHub GPT-5 codex models to responses transport', () => {
+  process.env.CLAUDE_CODE_USE_GITHUB = '1'
+  const r = resolveProviderRequest({ model: 'gpt-5.3-codex' })
+  expect(r.resolvedModel).toBe('gpt-5.3-codex')
+  expect(r.transport).toBe('codex_responses')
+})
+
+test('resolveProviderRequest keeps gpt-5-mini on chat_completions for GitHub', () => {
+  process.env.CLAUDE_CODE_USE_GITHUB = '1'
+  const r = resolveProviderRequest({ model: 'gpt-5-mini' })
+  expect(r.resolvedModel).toBe('gpt-5-mini')
+  expect(r.transport).toBe('chat_completions')
+})
+
 test('resolveProviderRequest leaves model unchanged without GitHub flag', () => {
   delete process.env.CLAUDE_CODE_USE_GITHUB
   const r = resolveProviderRequest({ model: 'github:gpt-4o' })

src/services/api/providerConfig.ts

@@ -7,8 +7,8 @@ import { isEnvTruthy } from '../../utils/envUtils.js'
 
 export const DEFAULT_OPENAI_BASE_URL = 'https://api.openai.com/v1'
 export const DEFAULT_CODEX_BASE_URL = 'https://chatgpt.com/backend-api/codex'
-/** Default GitHub Models API model when user selects copilot / github:copilot */
+/** Default GitHub Copilot API model when user selects copilot / github:copilot */
-export const DEFAULT_GITHUB_MODELS_API_MODEL = 'openai/gpt-4.1'
+export const DEFAULT_GITHUB_MODELS_API_MODEL = 'gpt-4o'
 
 const CODEX_ALIAS_MODELS: Record<
   string,
@@ -227,6 +227,21 @@ export function shouldUseCodexTransport(
   return isCodexBaseUrl(explicitBaseUrl) || (!explicitBaseUrl && isCodexAlias(model))
 }
 
+function shouldUseGithubResponsesApi(model: string): boolean {
+  const normalized = model.trim().toLowerCase()
+
+  // Codex-branded models require /responses.
+  if (normalized.includes('codex')) return true
+
+  // GPT-5+ models use /responses, except gpt-5-mini.
+  const match = /^gpt-(\d+)/.exec(normalized)
+  if (!match) return false
+  const major = Number(match[1])
+  if (major < 5) return false
+  if (normalized.startsWith('gpt-5-mini')) return false
+  return true
+}
+
 export function isLocalProviderUrl(baseUrl: string | undefined): boolean {
   if (!baseUrl) return false
   try {
@@ -280,19 +295,61 @@ export function isCodexBaseUrl(baseUrl: string | undefined): boolean {
 }
 
 /**
- * Normalize user model string for GitHub Models inference (models.github.ai).
+ * Normalize user model string for GitHub Copilot API inference.
- * Mirrors runtime devsper `github._normalize_model_id`.
+ * Mirrors how Copilot resolves model IDs internally.
  */
-export function normalizeGithubModelsApiModel(requestedModel: string): string {
+export function normalizeGithubCopilotModel(requestedModel: string): string {
   const noQuery = requestedModel.split('?', 1)[0] ?? requestedModel
   const segment =
     noQuery.includes(':') ? noQuery.split(':', 2)[1]!.trim() : noQuery.trim()
   if (!segment || segment.toLowerCase() === 'copilot') {
     return DEFAULT_GITHUB_MODELS_API_MODEL
   }
+  // Strip provider prefix if present (e.g., "openai/gpt-4o" -> "gpt-4o")
+  const slashIndex = segment.indexOf('/')
+  if (slashIndex !== -1) {
+    return segment.slice(slashIndex + 1)
+  }
   return segment
 }
 
+/**
+ * Normalize user model string for GitHub Models API inference.
+ * Only normalizes the default alias, preserves provider-qualified models.
+ */
+export function normalizeGithubModelsApiModel(requestedModel: string): string {
+  const noQuery = requestedModel.split('?', 1)[0] ?? requestedModel
+  const segment =
+    noQuery.includes(':') ? noQuery.split(':', 2)[1]!.trim() : noQuery.trim()
+  // Only normalize the default alias for GitHub Models
+  if (!segment || segment.toLowerCase() === 'copilot') {
+    return DEFAULT_GITHUB_MODELS_API_MODEL
+  }
+  // Preserve provider prefix for GitHub Models (e.g., "openai/gpt-4.1" stays as-is)
+  return segment
+}
+
+export const GITHUB_COPILOT_BASE_URL = 'https://api.githubcopilot.com'
+export const GITHUB_MODELS_BASE_URL = 'https://models.github.ai/inference'
+
+export function getGithubEndpointType(
+  baseUrl: string | undefined,
+): 'copilot' | 'models' | 'custom' {
+  if (!baseUrl) return 'copilot'
+  try {
+    const hostname = new URL(baseUrl).hostname.toLowerCase()
+    if (hostname === 'api.githubcopilot.com') {
+      return 'copilot'
+    }
+    if (hostname === 'models.github.ai' || hostname.endsWith('.github.ai')) {
+      return 'models'
+    }
+    return 'custom'
+  } catch {
+    return 'copilot'
+  }
+}
+
 export function resolveProviderRequest(options?: {
   model?: string
   baseUrl?: string
@@ -310,31 +367,49 @@ export function resolveProviderRequest(options?: {
     asEnvUrl(options?.baseUrl) ??
     asEnvUrl(process.env.OPENAI_BASE_URL) ??
     asEnvUrl(process.env.OPENAI_API_BASE)
+
+  const githubEndpointType = isGithubMode
+    ? getGithubEndpointType(rawBaseUrl)
+    : 'custom'
+  const isGithubCopilot = isGithubMode && githubEndpointType === 'copilot'
+  const isGithubModels = isGithubMode && githubEndpointType === 'models'
+  const isGithubCustom = isGithubMode && githubEndpointType === 'custom'
+
+  const githubResolvedModel = isGithubMode
+    ? normalizeGithubModelsApiModel(requestedModel)
+    : requestedModel
+
   const transport: ProviderTransport =
-    shouldUseCodexTransport(requestedModel, rawBaseUrl)
+    shouldUseCodexTransport(requestedModel, rawBaseUrl) ||
+      (isGithubCopilot && shouldUseGithubResponsesApi(githubResolvedModel))
       ? 'codex_responses'
       : 'chat_completions'
 
-  const resolvedModel =
+  // For GitHub Copilot API, normalize to real model ID (e.g., "github:copilot" -> "gpt-4o")
-    transport === 'chat_completions' &&
+  // For GitHub Models/custom endpoints:
-    isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB)
+  //   - Normalize default alias (github:copilot -> gpt-4o)
-      ? normalizeGithubModelsApiModel(requestedModel)
+  //   - Preserve provider-qualified models (openai/gpt-4.1 stays as-is)
-      : descriptor.baseModel
+  const resolvedModel = isGithubCopilot
+    ? normalizeGithubCopilotModel(descriptor.baseModel)
+    : (isGithubModels || isGithubCustom
+      ? normalizeGithubModelsApiModel(descriptor.baseModel)
+      : descriptor.baseModel)
 
   const reasoning = options?.reasoningEffortOverride
     ? { effort: options.reasoningEffortOverride }
     : descriptor.reasoning
 
-
   return {
     transport,
     requestedModel,
     resolvedModel,
     baseUrl:
       (rawBaseUrl ??
-        (transport === 'codex_responses'
+        (isGithubCopilot && transport === 'codex_responses'
-          ? DEFAULT_CODEX_BASE_URL
+          ? GITHUB_COPILOT_BASE_URL
-          : DEFAULT_OPENAI_BASE_URL)
+          : (isGithubMode
+            ? GITHUB_COPILOT_BASE_URL
+            : DEFAULT_OPENAI_BASE_URL))
       ).replace(/\/+$/, ''),
     reasoning,
   }

src/services/api/toolArgumentNormalization.test.ts

@@ -0,0 +1,180 @@
+import { describe, expect, test } from 'bun:test'
+import { normalizeToolArguments } from './toolArgumentNormalization'
+
+describe('normalizeToolArguments', () => {
+  describe('Bash tool', () => {
+    test('wraps plain string into { command }', () => {
+      expect(normalizeToolArguments('Bash', 'pwd')).toEqual({ command: 'pwd' })
+    })
+
+    test('wraps multi-word command', () => {
+      expect(normalizeToolArguments('Bash', 'ls -la /tmp')).toEqual({
+        command: 'ls -la /tmp',
+      })
+    })
+
+    test('passes through structured JSON object', () => {
+      expect(
+        normalizeToolArguments('Bash', '{"command":"echo hi"}'),
+      ).toEqual({ command: 'echo hi' })
+    })
+
+    test('returns empty object for blank string', () => {
+      expect(normalizeToolArguments('Bash', '')).toEqual({})
+      expect(normalizeToolArguments('Bash', '   ')).toEqual({})
+    })
+
+    test('returns parsed blank for JSON-encoded blank string', () => {
+      expect(normalizeToolArguments('Bash', '""')).toEqual('')
+      expect(normalizeToolArguments('Bash', '"  "')).toEqual('  ')
+    })
+
+    test('returns empty object for malformed structured object literal', () => {
+      expect(normalizeToolArguments('Bash', '{ "command": "pwd"')).toEqual({})
+    })
+
+    test.each([
+      ['{command:"pwd"}'],
+      ["{'command':'pwd'}"],
+      ['{command: pwd}'],
+    ])(
+      'returns empty object for malformed object-shaped string %s (does not wrap into command)',
+      (input) => {
+        expect(normalizeToolArguments('Bash', input)).toEqual({})
+      },
+    )
+
+    test.each([
+      ['false', false],
+      ['null', null],
+      ['[]', [] as unknown[]],
+      ['0', 0],
+      ['true', true],
+      ['123', 123],
+    ])(
+      'preserves JSON literal %s as-is (does not wrap into command)',
+      (input, expected) => {
+        expect(normalizeToolArguments('Bash', input)).toEqual(expected)
+      },
+    )
+
+    test('wraps JSON-encoded string into { command }', () => {
+      expect(normalizeToolArguments('Bash', '"pwd"')).toEqual({
+        command: 'pwd',
+      })
+    })
+  })
+
+  describe('undefined arguments', () => {
+    test('returns empty object for undefined', () => {
+      expect(normalizeToolArguments('Bash', undefined)).toEqual({})
+      expect(normalizeToolArguments('UnknownTool', undefined)).toEqual({})
+    })
+  })
+
+  describe('Read tool', () => {
+    test('wraps plain string into { file_path }', () => {
+      expect(normalizeToolArguments('Read', '/home/user/file.txt')).toEqual({
+        file_path: '/home/user/file.txt',
+      })
+    })
+
+    test('wraps JSON-encoded string into { file_path }', () => {
+      expect(normalizeToolArguments('Read', '"/home/user/file.txt"')).toEqual({
+        file_path: '/home/user/file.txt',
+      })
+    })
+
+    test('passes through structured JSON object', () => {
+      expect(
+        normalizeToolArguments('Read', '{"file_path":"/tmp/f.txt","limit":10}'),
+      ).toEqual({ file_path: '/tmp/f.txt', limit: 10 })
+    })
+  })
+
+  describe('Write tool', () => {
+    test('wraps plain string into { file_path }', () => {
+      expect(normalizeToolArguments('Write', '/tmp/out.txt')).toEqual({
+        file_path: '/tmp/out.txt',
+      })
+    })
+
+    test('passes through structured JSON object', () => {
+      expect(
+        normalizeToolArguments(
+          'Write',
+          '{"file_path":"/tmp/out.txt","content":"hello"}',
+        ),
+      ).toEqual({ file_path: '/tmp/out.txt', content: 'hello' })
+    })
+  })
+
+  describe('Edit tool', () => {
+    test('wraps plain string into { file_path }', () => {
+      expect(normalizeToolArguments('Edit', '/tmp/edit.ts')).toEqual({
+        file_path: '/tmp/edit.ts',
+      })
+    })
+
+    test('passes through structured JSON object', () => {
+      expect(
+        normalizeToolArguments(
+          'Edit',
+          '{"file_path":"/tmp/f.ts","old_string":"a","new_string":"b"}',
+        ),
+      ).toEqual({ file_path: '/tmp/f.ts', old_string: 'a', new_string: 'b' })
+    })
+  })
+
+  describe('Glob tool', () => {
+    test('wraps plain string into { pattern }', () => {
+      expect(normalizeToolArguments('Glob', '**/*.ts')).toEqual({
+        pattern: '**/*.ts',
+      })
+    })
+
+    test('passes through structured JSON object', () => {
+      expect(
+        normalizeToolArguments('Glob', '{"pattern":"*.js","path":"/src"}'),
+      ).toEqual({ pattern: '*.js', path: '/src' })
+    })
+  })
+
+  describe('Grep tool', () => {
+    test('wraps plain string into { pattern }', () => {
+      expect(normalizeToolArguments('Grep', 'TODO')).toEqual({
+        pattern: 'TODO',
+      })
+    })
+
+    test('passes through structured JSON object', () => {
+      expect(
+        normalizeToolArguments('Grep', '{"pattern":"fixme","path":"/src"}'),
+      ).toEqual({ pattern: 'fixme', path: '/src' })
+    })
+  })
+
+  describe('unknown tools', () => {
+    test('returns empty object for plain string (no known field mapping)', () => {
+      expect(normalizeToolArguments('UnknownTool', 'some value')).toEqual({})
+    })
+
+    test('passes through structured JSON object', () => {
+      expect(
+        normalizeToolArguments('UnknownTool', '{"key":"val"}'),
+      ).toEqual({ key: 'val' })
+    })
+
+    test('preserves JSON literals as-is', () => {
+      expect(normalizeToolArguments('UnknownTool', 'false')).toEqual(false)
+      expect(normalizeToolArguments('UnknownTool', 'null')).toEqual(null)
+      expect(normalizeToolArguments('UnknownTool', '[]')).toEqual([])
+    })
+
+    test('returns parsed string for JSON-encoded string on unknown tools', () => {
+      expect(normalizeToolArguments('UnknownTool', '"hello"')).toEqual(
+        'hello',
+      )
+    })
+  })
+})

src/services/api/toolArgumentNormalization.ts

@@ -0,0 +1,69 @@
+const STRING_ARGUMENT_TOOL_FIELDS: Record<string, string> = {
+  Bash: 'command',
+  Read: 'file_path',
+  Write: 'file_path',
+  Edit: 'file_path',
+  Glob: 'pattern',
+  Grep: 'pattern',
+}
+
+function isBlankString(value: string): boolean {
+  return value.trim().length === 0
+}
+
+function isLikelyStructuredObjectLiteral(value: string): boolean {
+  // Match object-like patterns with key-value syntax:
+  // {"key":, {key:, {'key':, { "key" :, etc.
+  // But NOT bash compound commands like { pwd; } or { echo hi; }
+  return /^\s*\{\s*['"]?\w+['"]?\s*:/.test(value)
+}
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === 'object' && value !== null && !Array.isArray(value)
+}
+
+function getPlainStringToolArgumentField(toolName: string): string | null {
+  return STRING_ARGUMENT_TOOL_FIELDS[toolName] ?? null
+}
+
+export function hasToolFieldMapping(toolName: string): boolean {
+  return toolName in STRING_ARGUMENT_TOOL_FIELDS
+}
+
+function wrapPlainStringToolArguments(
+  toolName: string,
+  value: string,
+): Record<string, string> | null {
+  const field = getPlainStringToolArgumentField(toolName)
+  if (!field) return null
+  return { [field]: value }
+}
+
+export function normalizeToolArguments(
+  toolName: string,
+  rawArguments: string | undefined,
+): unknown {
+  if (rawArguments === undefined) return {}
+
+  try {
+    const parsed = JSON.parse(rawArguments)
+    if (isRecord(parsed)) {
+      return parsed
+    }
+    // Parsed as a non-object JSON value (string, number, boolean, null, array)
+    if (typeof parsed === 'string' && !isBlankString(parsed)) {
+      return wrapPlainStringToolArguments(toolName, parsed) ?? parsed
+    }
+    // For blank strings, booleans, null, arrays — pass through as-is
+    // and let Zod schema validation produce a meaningful error
+    return parsed
+  } catch {
+    // rawArguments is not valid JSON — treat as a plain string
+    if (isBlankString(rawArguments) || isLikelyStructuredObjectLiteral(rawArguments)) {
+      // Blank or looks like a malformed object literal — don't wrap into
+      // a tool field to avoid turning garbage into executable input
+      return {}
+    }
+    return wrapPlainStringToolArguments(toolName, rawArguments) ?? {}
+  }
+}

src/services/api/withRetry.test.ts

@@ -1,4 +1,4 @@
-import { afterEach, describe, expect, mock, test } from 'bun:test'
+import { afterEach, beforeEach, describe, expect, mock, test } from 'bun:test'
 import { APIError } from '@anthropic-ai/sdk'
 
 // Helper to build a mock APIError with specific headers
@@ -15,15 +15,27 @@ function makeError(headers: Record<string, string>): APIError {
 
 // Save/restore env vars between tests
 const originalEnv = { ...process.env }
-afterEach(() => {
+
-  for (const key of [
+const envKeys = [
   'CLAUDE_CODE_USE_OPENAI',
   'CLAUDE_CODE_USE_GEMINI',
   'CLAUDE_CODE_USE_GITHUB',
   'CLAUDE_CODE_USE_BEDROCK',
   'CLAUDE_CODE_USE_VERTEX',
   'CLAUDE_CODE_USE_FOUNDRY',
-  ]) {
+  'OPENAI_MODEL',
+  'OPENAI_BASE_URL',
+  'OPENAI_API_BASE',
+] as const
+
+beforeEach(() => {
+  for (const key of envKeys) {
+    delete process.env[key]
+  }
+})
+
+afterEach(() => {
+  for (const key of envKeys) {
     if (originalEnv[key] === undefined) delete process.env[key]
     else process.env[key] = originalEnv[key]
   }

src/services/autoFix/autoFixConfig.test.ts

@@ -0,0 +1,106 @@
+import { describe, expect, test } from 'bun:test'
+import { AutoFixConfigSchema, getAutoFixConfig, type AutoFixConfig } from './autoFixConfig.js'
+
+describe('AutoFixConfigSchema', () => {
+  test('parses valid full config', () => {
+    const input = {
+      enabled: true,
+      lint: 'eslint . --fix',
+      test: 'bun test',
+      maxRetries: 3,
+      timeout: 30000,
+    }
+    const result = AutoFixConfigSchema.safeParse(input)
+    expect(result.success).toBe(true)
+    if (result.success) {
+      expect(result.data.enabled).toBe(true)
+      expect(result.data.lint).toBe('eslint . --fix')
+      expect(result.data.test).toBe('bun test')
+      expect(result.data.maxRetries).toBe(3)
+      expect(result.data.timeout).toBe(30000)
+    }
+  })
+
+  test('parses minimal config with defaults', () => {
+    const input = { enabled: true, lint: 'eslint .' }
+    const result = AutoFixConfigSchema.safeParse(input)
+    expect(result.success).toBe(true)
+    if (result.success) {
+      expect(result.data.maxRetries).toBe(3)
+      expect(result.data.timeout).toBe(30000)
+      expect(result.data.test).toBeUndefined()
+    }
+  })
+
+  test('rejects config with enabled but no lint or test', () => {
+    const input = { enabled: true }
+    const result = AutoFixConfigSchema.safeParse(input)
+    expect(result.success).toBe(false)
+  })
+
+  test('accepts disabled config without commands', () => {
+    const input = { enabled: false }
+    const result = AutoFixConfigSchema.safeParse(input)
+    expect(result.success).toBe(true)
+  })
+
+  test('rejects negative maxRetries', () => {
+    const input = { enabled: true, lint: 'eslint .', maxRetries: -1 }
+    const result = AutoFixConfigSchema.safeParse(input)
+    expect(result.success).toBe(false)
+  })
+
+  test('rejects maxRetries above 10', () => {
+    const input = { enabled: true, lint: 'eslint .', maxRetries: 11 }
+    const result = AutoFixConfigSchema.safeParse(input)
+    expect(result.success).toBe(false)
+  })
+})
+
+describe('getAutoFixConfig', () => {
+  test('returns null when settings have no autoFix', () => {
+    const result = getAutoFixConfig(undefined)
+    expect(result).toBeNull()
+  })
+
+  test('returns null when autoFix is disabled', () => {
+    const result = getAutoFixConfig({ enabled: false })
+    expect(result).toBeNull()
+  })
+
+  test('returns parsed config when valid and enabled', () => {
+    const result = getAutoFixConfig({ enabled: true, lint: 'eslint .' })
+    expect(result).not.toBeNull()
+    expect(result!.enabled).toBe(true)
+    expect(result!.lint).toBe('eslint .')
+  })
+})
+
+describe('SettingsSchema autoFix integration', () => {
+  test('SettingsSchema accepts autoFix field', async () => {
+    const { SettingsSchema } = await import('../../utils/settings/types.js')
+    const settings = {
+      autoFix: {
+        enabled: true,
+        lint: 'eslint .',
+        test: 'bun test',
+        maxRetries: 3,
+        timeout: 30000,
+      },
+    }
+    const result = SettingsSchema().safeParse(settings)
+    expect(result.success).toBe(true)
+  })
+
+  test('SettingsSchema rejects invalid autoFix', async () => {
+    const { SettingsSchema } = await import('../../utils/settings/types.js')
+    const settings = {
+      autoFix: {
+        enabled: true,
+        // missing lint and test - should fail refine
+      },
+    }
+    const result = SettingsSchema().safeParse(settings)
+    expect(result.success).toBe(false)
+  })
+})

src/services/autoFix/autoFixConfig.ts

@@ -0,0 +1,52 @@
+import { z } from 'zod/v4'
+
+export const AutoFixConfigSchema = z
+  .object({
+    enabled: z.boolean().describe('Whether auto-fix is enabled'),
+    lint: z
+      .string()
+      .optional()
+      .describe('Lint command to run after file edits (e.g. "eslint . --fix")'),
+    test: z
+      .string()
+      .optional()
+      .describe('Test command to run after file edits (e.g. "bun test")'),
+    maxRetries: z
+      .number()
+      .int()
+      .min(0)
+      .max(10)
+      .default(3)
+      .describe('Maximum number of auto-fix retry attempts (default: 3)'),
+    timeout: z
+      .number()
+      .int()
+      .min(1000)
+      .max(300000)
+      .default(30000)
+      .describe('Timeout in ms for each lint/test command (default: 30000)'),
+  })
+  .refine(
+    data => !data.enabled || data.lint !== undefined || data.test !== undefined,
+    {
+      message: 'At least one of "lint" or "test" must be set when enabled',
+    },
+  )
+
+export type AutoFixConfig = z.infer<typeof AutoFixConfigSchema>
+
+export function getAutoFixConfig(
+  rawConfig: unknown,
+): AutoFixConfig | null {
+  if (!rawConfig || typeof rawConfig !== 'object') {
+    return null
+  }
+  const parsed = AutoFixConfigSchema.safeParse(rawConfig)
+  if (!parsed.success) {
+    return null
+  }
+  if (!parsed.data.enabled) {
+    return null
+  }
+  return parsed.data
+}

src/services/autoFix/autoFixHook.test.ts

@@ -0,0 +1,63 @@
+import { describe, expect, test } from 'bun:test'
+import {
+  shouldRunAutoFix,
+  buildAutoFixContext,
+} from './autoFixHook.js'
+
+describe('shouldRunAutoFix', () => {
+  test('returns true for file_edit tool when autoFix enabled', () => {
+    const config = { enabled: true, lint: 'eslint .', maxRetries: 3, timeout: 30000 }
+    expect(shouldRunAutoFix('file_edit', config)).toBe(true)
+  })
+
+  test('returns true for file_write tool when autoFix enabled', () => {
+    const config = { enabled: true, lint: 'eslint .', maxRetries: 3, timeout: 30000 }
+    expect(shouldRunAutoFix('file_write', config)).toBe(true)
+  })
+
+  test('returns false for bash tool', () => {
+    const config = { enabled: true, lint: 'eslint .', maxRetries: 3, timeout: 30000 }
+    expect(shouldRunAutoFix('bash', config)).toBe(false)
+  })
+
+  test('returns false for file_read tool', () => {
+    const config = { enabled: true, lint: 'eslint .', maxRetries: 3, timeout: 30000 }
+    expect(shouldRunAutoFix('file_read', config)).toBe(false)
+  })
+
+  test('returns false when config is null', () => {
+    expect(shouldRunAutoFix('file_edit', null)).toBe(false)
+  })
+})
+
+describe('buildAutoFixContext', () => {
+  test('formats lint errors as AI-readable context', () => {
+    const context = buildAutoFixContext({
+      hasErrors: true,
+      lintOutput: 'src/foo.ts:10:5 error no-unused-vars',
+      lintExitCode: 1,
+      errorSummary: 'Lint errors (exit code 1):\nsrc/foo.ts:10:5 error no-unused-vars',
+    })
+    expect(context).toContain('AUTO-FIX')
+    expect(context).toContain('no-unused-vars')
+    expect(context).toContain('Please fix')
+  })
+
+  test('returns null when no errors', () => {
+    const context = buildAutoFixContext({
+      hasErrors: false,
+    })
+    expect(context).toBeNull()
+  })
+
+  test('formats test failures as AI-readable context', () => {
+    const context = buildAutoFixContext({
+      hasErrors: true,
+      testOutput: 'FAIL src/foo.test.ts\n  expected true, got false',
+      testExitCode: 1,
+      errorSummary: 'Test failures (exit code 1):\nFAIL src/foo.test.ts',
+    })
+    expect(context).toContain('AUTO-FIX')
+    expect(context).toContain('FAIL')
+  })
+})

src/services/autoFix/autoFixHook.ts

@@ -0,0 +1,25 @@
+import type { AutoFixConfig } from './autoFixConfig.js'
+import type { AutoFixResult } from './autoFixRunner.js'
+
+const AUTO_FIX_TOOLS = new Set(['file_edit', 'file_write'])
+
+export function shouldRunAutoFix(
+  toolName: string,
+  config: AutoFixConfig | null,
+): boolean {
+  if (!config) return false
+  return AUTO_FIX_TOOLS.has(toolName)
+}
+
+export function buildAutoFixContext(result: AutoFixResult): string | null {
+  if (!result.hasErrors || !result.errorSummary) return null
+
+  return (
+    `<auto_fix_feedback>\n` +
+    `AUTO-FIX: The file you just edited has errors. Please fix them:\n\n` +
+    `${result.errorSummary}\n\n` +
+    `Please fix these errors in the files you just edited. ` +
+    `Do not ask the user — just apply the fix.\n` +
+    `</auto_fix_feedback>`
+  )
+}

src/services/autoFix/autoFixIntegration.test.ts

@@ -0,0 +1,48 @@
+import { describe, expect, test } from 'bun:test'
+import { getAutoFixConfig } from './autoFixConfig.js'
+import { shouldRunAutoFix, buildAutoFixContext } from './autoFixHook.js'
+import { runAutoFixCheck } from './autoFixRunner.js'
+
+describe('autoFix end-to-end flow', () => {
+  test('full flow: config → shouldRun → check → context', async () => {
+    const config = getAutoFixConfig({
+      enabled: true,
+      lint: 'echo "error: unused" && exit 1',
+      maxRetries: 2,
+      timeout: 5000,
+    })
+    expect(config).not.toBeNull()
+    expect(shouldRunAutoFix('file_edit', config)).toBe(true)
+
+    const result = await runAutoFixCheck({
+      lint: config!.lint,
+      test: config!.test,
+      timeout: config!.timeout,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(true)
+
+    const context = buildAutoFixContext(result)
+    expect(context).not.toBeNull()
+    expect(context).toContain('AUTO-FIX')
+    expect(context).toContain('unused')
+  })
+
+  test('full flow: no errors = no context', async () => {
+    const config = getAutoFixConfig({
+      enabled: true,
+      lint: 'echo "all clean"',
+      timeout: 5000,
+    })
+    const result = await runAutoFixCheck({
+      lint: config!.lint,
+      timeout: config!.timeout,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(false)
+    const context = buildAutoFixContext(result)
+    expect(context).toBeNull()
+  })
+})

src/services/autoFix/autoFixRunner.test.ts

@@ -0,0 +1,103 @@
+import { describe, expect, test } from 'bun:test'
+import {
+  runAutoFixCheck,
+  type AutoFixResult,
+  type AutoFixCheckOptions,
+} from './autoFixRunner.js'
+
+describe('runAutoFixCheck', () => {
+  test('returns success when lint command exits 0', async () => {
+    const result = await runAutoFixCheck({
+      lint: 'echo "all clean"',
+      timeout: 5000,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(false)
+    expect(result.lintOutput).toContain('all clean')
+    expect(result.testOutput).toBeUndefined()
+  })
+
+  test('returns errors when lint command exits non-zero', async () => {
+    const result = await runAutoFixCheck({
+      lint: 'echo "error: unused var" && exit 1',
+      timeout: 5000,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(true)
+    expect(result.lintOutput).toContain('unused var')
+    expect(result.lintExitCode).toBe(1)
+  })
+
+  test('returns errors when test command exits non-zero', async () => {
+    const result = await runAutoFixCheck({
+      test: 'echo "FAIL test_foo" && exit 1',
+      timeout: 5000,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(true)
+    expect(result.testOutput).toContain('FAIL test_foo')
+    expect(result.testExitCode).toBe(1)
+  })
+
+  test('runs both lint and test commands', async () => {
+    const result = await runAutoFixCheck({
+      lint: 'echo "lint ok"',
+      test: 'echo "test ok"',
+      timeout: 5000,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(false)
+    expect(result.lintOutput).toContain('lint ok')
+    expect(result.testOutput).toContain('test ok')
+  })
+
+  test('skips test if lint fails', async () => {
+    const result = await runAutoFixCheck({
+      lint: 'echo "lint error" && exit 1',
+      test: 'echo "should not run"',
+      timeout: 5000,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(true)
+    expect(result.lintOutput).toContain('lint error')
+    expect(result.testOutput).toBeUndefined()
+  })
+
+  test('handles timeout gracefully', async () => {
+    const result = await runAutoFixCheck({
+      lint: 'sleep 10',
+      timeout: 100,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(true)
+    expect(result.timedOut).toBe(true)
+  })
+
+  test('returns success with no commands configured', async () => {
+    const result = await runAutoFixCheck({
+      timeout: 5000,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(false)
+  })
+
+  test('formats error summary for AI consumption', async () => {
+    const result = await runAutoFixCheck({
+      lint: 'echo "src/foo.ts:10:5 error no-unused-vars" && exit 1',
+      timeout: 5000,
+
+      cwd: '/tmp',
+    })
+    expect(result.hasErrors).toBe(true)
+    const summary = result.errorSummary
+    expect(summary).toContain('Lint errors')
+    expect(summary).toContain('no-unused-vars')
+  })
+})

src/services/autoFix/autoFixRunner.ts

@@ -0,0 +1,169 @@
+import { spawn } from 'child_process'
+
+export interface AutoFixCheckOptions {
+  lint?: string
+  test?: string
+  timeout: number
+  cwd: string
+  signal?: AbortSignal
+}
+
+export interface AutoFixResult {
+  hasErrors: boolean
+  lintOutput?: string
+  lintExitCode?: number
+  testOutput?: string
+  testExitCode?: number
+  timedOut?: boolean
+  errorSummary?: string
+}
+
+async function runCommand(
+  command: string,
+  cwd: string,
+  timeout: number,
+  signal?: AbortSignal,
+): Promise<{ stdout: string; stderr: string; exitCode: number; timedOut: boolean }> {
+  return new Promise((resolve) => {
+    if (signal?.aborted) {
+      resolve({ stdout: '', stderr: 'Aborted', exitCode: 1, timedOut: false })
+      return
+    }
+
+    let timedOut = false
+    let stdout = ''
+    let stderr = ''
+
+    const isWindows = process.platform === 'win32'
+    const proc = spawn(command, [], {
+      cwd,
+      env: { ...process.env },
+      shell: true,
+      windowsHide: true,
+      // On Unix, create a process group so we can kill child processes on timeout/abort
+      detached: !isWindows,
+    })
+
+    const killTree = () => {
+      try {
+        if (!isWindows && proc.pid) {
+          // Kill the entire process group
+          process.kill(-proc.pid, 'SIGTERM')
+        } else {
+          proc.kill('SIGTERM')
+        }
+      } catch {
+        // Process may have already exited
+      }
+    }
+
+    const onAbort = () => {
+      killTree()
+    }
+    signal?.addEventListener('abort', onAbort, { once: true })
+
+    proc.stdout?.on('data', (data: Buffer) => {
+      stdout += data.toString()
+    })
+    proc.stderr?.on('data', (data: Buffer) => {
+      stderr += data.toString()
+    })
+
+    const timer = setTimeout(() => {
+      timedOut = true
+      killTree()
+    }, timeout)
+
+    proc.on('close', (code) => {
+      clearTimeout(timer)
+      signal?.removeEventListener('abort', onAbort)
+      resolve({
+        stdout: stdout.slice(0, 10000),
+        stderr: stderr.slice(0, 10000),
+        exitCode: code ?? 1,
+        timedOut,
+      })
+    })
+
+    proc.on('error', () => {
+      clearTimeout(timer)
+      signal?.removeEventListener('abort', onAbort)
+      resolve({
+        stdout,
+        stderr: stderr || 'Command failed to start',
+        exitCode: 1,
+        timedOut: false,
+      })
+    })
+  })
+}
+
+function buildErrorSummary(result: AutoFixResult): string | undefined {
+  if (!result.hasErrors) return undefined
+  const parts: string[] = []
+
+  if (result.timedOut) {
+    parts.push('Command timed out.')
+  }
+  if (result.lintExitCode !== undefined && result.lintExitCode !== 0) {
+    parts.push(`Lint errors (exit code ${result.lintExitCode}):\n${result.lintOutput ?? ''}`)
+  }
+  if (result.testExitCode !== undefined && result.testExitCode !== 0) {
+    parts.push(`Test failures (exit code ${result.testExitCode}):\n${result.testOutput ?? ''}`)
+  }
+
+  return parts.join('\n\n')
+}
+
+export async function runAutoFixCheck(
+  options: AutoFixCheckOptions,
+): Promise<AutoFixResult> {
+  const { lint, test, timeout, cwd, signal } = options
+
+  if (!lint && !test) {
+    return { hasErrors: false }
+  }
+
+  if (signal?.aborted) {
+    return { hasErrors: false }
+  }
+
+  const result: AutoFixResult = { hasErrors: false }
+
+  // Run lint first
+  if (lint) {
+    const lintResult = await runCommand(lint, cwd, timeout, signal)
+    result.lintOutput = (lintResult.stdout + '\n' + lintResult.stderr).trim()
+    result.lintExitCode = lintResult.exitCode
+
+    if (lintResult.timedOut) {
+      result.hasErrors = true
+      result.timedOut = true
+      result.errorSummary = buildErrorSummary(result)
+      return result
+    }
+
+    if (lintResult.exitCode !== 0) {
+      result.hasErrors = true
+      result.errorSummary = buildErrorSummary(result)
+      return result
+    }
+  }
+
+  // Run tests only if lint passed (or no lint configured)
+  if (test) {
+    const testResult = await runCommand(test, cwd, timeout, signal)
+    result.testOutput = (testResult.stdout + '\n' + testResult.stderr).trim()
+    result.testExitCode = testResult.exitCode
+
+    if (testResult.timedOut) {
+      result.hasErrors = true
+      result.timedOut = true
+    } else if (testResult.exitCode !== 0) {
+      result.hasErrors = true
+    }
+  }
+
+  result.errorSummary = buildErrorSummary(result)
+  return result
+}

src/services/compact/microCompact.test.ts

@@ -0,0 +1,127 @@
+import { describe, expect, test } from 'bun:test'
+
+import type { Message } from '../../types/message.js'
+import { createAssistantMessage, createUserMessage } from '../../utils/messages.js'
+
+// We test the exported collectCompactableToolIds behavior indirectly via
+// the public microcompactMessages + time-based path. But first we need to
+// verify the core predicate: MCP tools (prefixed 'mcp__') should be
+// compactable alongside the built-in tool set.
+
+// Import internals we can test
+import { evaluateTimeBasedTrigger } from './microCompact.js'
+
+/**
+ * Helper: build a minimal assistant message with a tool_use block.
+ */
+function assistantWithToolUse(toolName: string, toolId: string): Message {
+  return createAssistantMessage({
+    content: [
+      {
+        type: 'tool_use' as const,
+        id: toolId,
+        name: toolName,
+        input: {},
+      },
+    ],
+  })
+}
+
+/**
+ * Helper: build a user message with a tool_result block.
+ */
+function userWithToolResult(toolId: string, output: string): Message {
+  return createUserMessage({
+    content: [
+      {
+        type: 'tool_result' as const,
+        tool_use_id: toolId,
+        content: output,
+      },
+    ],
+  })
+}
+
+describe('microCompact MCP tool compaction', () => {
+  // We can't easily unit-test the private isCompactableTool directly,
+  // but we can test the full time-based microcompact path which exercises
+  // collectCompactableToolIds → isCompactableTool under the hood.
+  // The time-based path is the simplest to trigger: it content-clears
+  // old tool results when the gap since last assistant message exceeds
+  // the threshold.
+
+  // However, evaluateTimeBasedTrigger depends on config (GrowthBook).
+  // So instead, let's test the observable behavior by importing the
+  // microcompactMessages function and checking that MCP tool_use blocks
+  // are collected.
+
+  // Since collectCompactableToolIds is not exported, we test the predicate
+  // behavior by verifying that the module loads without error and that
+  // built-in and MCP tools are treated consistently.
+
+  test('module exports load correctly', async () => {
+    const mod = await import('./microCompact.js')
+    expect(mod.microcompactMessages).toBeFunction()
+    expect(mod.estimateMessageTokens).toBeFunction()
+    expect(mod.evaluateTimeBasedTrigger).toBeFunction()
+  })
+
+  test('estimateMessageTokens counts MCP tool_use blocks', async () => {
+    const { estimateMessageTokens } = await import('./microCompact.js')
+
+    const builtinMessages: Message[] = [
+      assistantWithToolUse('Read', 'tool-builtin-1'),
+      userWithToolResult('tool-builtin-1', 'file contents here'),
+    ]
+
+    const mcpMessages: Message[] = [
+      assistantWithToolUse('mcp__github__get_file_contents', 'tool-mcp-1'),
+      userWithToolResult('tool-mcp-1', 'file contents here'),
+    ]
+
+    const builtinTokens = estimateMessageTokens(builtinMessages)
+    const mcpTokens = estimateMessageTokens(mcpMessages)
+
+    // Both should produce non-zero estimates
+    expect(builtinTokens).toBeGreaterThan(0)
+    expect(mcpTokens).toBeGreaterThan(0)
+
+    // The tool_result content is identical, so token estimates should be
+    // similar (tool_use name differs slightly, so not exactly equal)
+    expect(Math.abs(builtinTokens - mcpTokens)).toBeLessThan(50)
+  })
+
+  test('microcompactMessages processes MCP tools without error', async () => {
+    const { microcompactMessages } = await import('./microCompact.js')
+
+    const messages: Message[] = [
+      assistantWithToolUse('mcp__slack__send_message', 'tool-mcp-2'),
+      userWithToolResult('tool-mcp-2', 'Message sent successfully'),
+      assistantWithToolUse('mcp__github__create_pull_request', 'tool-mcp-3'),
+      userWithToolResult('tool-mcp-3', JSON.stringify({ number: 42, url: 'https://github.com/org/repo/pull/42' })),
+    ]
+
+    // Should not throw — MCP tools should be handled gracefully
+    const result = await microcompactMessages(messages)
+    expect(result).toBeDefined()
+    expect(result.messages).toBeDefined()
+    expect(result.messages.length).toBe(messages.length)
+  })
+
+  test('microcompactMessages processes mixed built-in and MCP tools', async () => {
+    const { microcompactMessages } = await import('./microCompact.js')
+
+    const messages: Message[] = [
+      assistantWithToolUse('Read', 'tool-read-1'),
+      userWithToolResult('tool-read-1', 'some file content'),
+      assistantWithToolUse('mcp__playwright__screenshot', 'tool-mcp-4'),
+      userWithToolResult('tool-mcp-4', 'base64-encoded-screenshot-data'.repeat(100)),
+      assistantWithToolUse('Bash', 'tool-bash-1'),
+      userWithToolResult('tool-bash-1', 'command output'),
+    ]
+
+    const result = await microcompactMessages(messages)
+    expect(result).toBeDefined()
+    expect(result.messages.length).toBe(messages.length)
+  })
+})

src/services/compact/microCompact.ts

@@ -37,7 +37,7 @@ export const TIME_BASED_MC_CLEARED_MESSAGE = '[Old tool result content cleared]'
 
 const IMAGE_MAX_TOKEN_SIZE = 2000
 
-// Only compact these tools
+// Only compact these built-in tools (MCP tools are also compactable via prefix match)
 const COMPACTABLE_TOOLS = new Set<string>([
   FILE_READ_TOOL_NAME,
   ...SHELL_TOOL_NAMES,
@@ -49,7 +49,13 @@ const COMPACTABLE_TOOLS = new Set<string>([
   FILE_WRITE_TOOL_NAME,
 ])
 
-// --- Cached microcompact state (internal-only, gated by feature('CACHED_MICROCOMPACT')) ---
+const MCP_TOOL_PREFIX = 'mcp__'
+
+function isCompactableTool(name: string): boolean {
+  return COMPACTABLE_TOOLS.has(name) || name.startsWith(MCP_TOOL_PREFIX)
+}
+
+// --- Cached microcompact state (gated by feature('CACHED_MICROCOMPACT')) ---
 
 // Lazy-initialized cached MC module and state to avoid importing in external builds.
 // The imports and state live inside feature() checks for dead code elimination.
@@ -231,7 +237,7 @@ function collectCompactableToolIds(messages: Message[]): string[] {
       Array.isArray(message.message.content)
     ) {
       for (const block of message.message.content) {
-        if (block.type === 'tool_use' && COMPACTABLE_TOOLS.has(block.name)) {
+        if (block.type === 'tool_use' && isCompactableTool(block.name)) {
           ids.push(block.id)
         }
       }

src/services/github/deviceFlow.test.ts

@@ -1,4 +1,4 @@
-import { afterEach, describe, expect, mock, test } from 'bun:test'
+import { afterEach, beforeEach, describe, expect, mock, test } from 'bun:test'
 
 import {
   DEFAULT_GITHUB_DEVICE_SCOPE,
@@ -7,14 +7,26 @@ import {
   requestDeviceCode,
 } from './deviceFlow.js'
 
+async function importFreshModule() {
+  mock.restore()
+  return import(`./deviceFlow.ts?ts=${Date.now()}-${Math.random()}`)
+}
+
 describe('requestDeviceCode', () => {
   const originalFetch = globalThis.fetch
 
+  beforeEach(() => {
+    mock.restore()
+    globalThis.fetch = originalFetch
+  })
+
   afterEach(() => {
     globalThis.fetch = originalFetch
   })
 
   test('parses successful device code response', async () => {
+    const { requestDeviceCode } = await importFreshModule()
+
     globalThis.fetch = mock(() =>
       Promise.resolve(
         new Response(
@@ -42,6 +54,9 @@ describe('requestDeviceCode', () => {
   })
 
   test('throws on HTTP error', async () => {
+    const { requestDeviceCode, GitHubDeviceFlowError } =
+      await importFreshModule()
+
     globalThis.fetch = mock(() =>
       Promise.resolve(new Response('bad', { status: 500 })),
     )
@@ -134,6 +149,8 @@ describe('pollAccessToken', () => {
   })
 
   test('returns token when GitHub responds with access_token immediately', async () => {
+    const { pollAccessToken } = await importFreshModule()
+
     let calls = 0
     globalThis.fetch = mock(() => {
       calls++
@@ -153,6 +170,8 @@ describe('pollAccessToken', () => {
   })
 
   test('throws on access_denied', async () => {
+    const { pollAccessToken } = await importFreshModule()
+
     globalThis.fetch = mock(() =>
       Promise.resolve(
         new Response(JSON.stringify({ error: 'access_denied' }), {
@@ -168,3 +187,62 @@ describe('pollAccessToken', () => {
     ).rejects.toThrow(/denied/)
   })
 })
+
+describe('exchangeForCopilotToken', () => {
+  const originalFetch = globalThis.fetch
+
+  afterEach(() => {
+    globalThis.fetch = originalFetch
+  })
+
+  test('parses successful Copilot token response', async () => {
+    const { exchangeForCopilotToken } = await importFreshModule()
+
+    globalThis.fetch = mock(() =>
+      Promise.resolve(
+        new Response(
+          JSON.stringify({
+            token: 'copilot-token-xyz',
+            expires_at: 1700000000,
+            refresh_in: 3600,
+            endpoints: {
+              api: 'https://api.githubcopilot.com',
+            },
+          }),
+          { status: 200 },
+        ),
+      ),
+    )
+
+    const result = await exchangeForCopilotToken('oauth-token', globalThis.fetch)
+    expect(result.token).toBe('copilot-token-xyz')
+    expect(result.expires_at).toBe(1700000000)
+    expect(result.refresh_in).toBe(3600)
+    expect(result.endpoints.api).toBe('https://api.githubcopilot.com')
+  })
+
+  test('throws on HTTP error', async () => {
+    const { exchangeForCopilotToken, GitHubDeviceFlowError } =
+      await importFreshModule()
+
+    globalThis.fetch = mock(() =>
+      Promise.resolve(new Response('unauthorized', { status: 401 })),
+    )
+    await expect(
+      exchangeForCopilotToken('bad-token', globalThis.fetch),
+    ).rejects.toThrow(GitHubDeviceFlowError)
+  })
+
+  test('throws on malformed response', async () => {
+    const { exchangeForCopilotToken } = await importFreshModule()
+
+    globalThis.fetch = mock(() =>
+      Promise.resolve(
+        new Response(JSON.stringify({ invalid: 'data' }), { status: 200 }),
+      ),
+    )
+    await expect(
+      exchangeForCopilotToken('oauth-token', globalThis.fetch),
+    ).rejects.toThrow(/Malformed/)
+  })
+})

src/services/github/deviceFlow.ts

@@ -1,19 +1,35 @@
 /**
  * GitHub OAuth device flow for CLI login (https://docs.github.com/en/apps/oauth-apps/building-oauth-apps/authorizing-oauth-apps#device-flow).
+ * Uses GitHub Copilot's official OAuth app for device authentication.
  */
 
 import { execFileNoThrow } from '../../utils/execFileNoThrow.js'
 
-export const DEFAULT_GITHUB_DEVICE_FLOW_CLIENT_ID = 'Ov23liXjWSSui6QIahPl'
+export const DEFAULT_GITHUB_DEVICE_FLOW_CLIENT_ID = 'Iv1.b507a08c87ecfe98'
 
 export const GITHUB_DEVICE_CODE_URL = 'https://github.com/login/device/code'
 export const GITHUB_DEVICE_ACCESS_TOKEN_URL =
   'https://github.com/login/oauth/access_token'
+export const COPILOT_TOKEN_URL = 'https://api.github.com/copilot_internal/v2/token'
 
-// OAuth app device flow does not accept the GitHub Models permission token
+/** Only read:user scope — required for Copilot OAuth */
-// scope (models:read). Use an OAuth-safe default.
+export const DEFAULT_GITHUB_DEVICE_SCOPE = 'read:user'
-const OAUTH_SAFE_GITHUB_DEVICE_SCOPE = 'read:user'
+
-export const DEFAULT_GITHUB_DEVICE_SCOPE = OAUTH_SAFE_GITHUB_DEVICE_SCOPE
+export const COPILOT_HEADERS: Record<string, string> = {
+  'User-Agent': 'GitHubCopilotChat/0.26.7',
+  'Editor-Version': 'vscode/1.99.3',
+  'Editor-Plugin-Version': 'copilot-chat/0.26.7',
+  'Copilot-Integration-Id': 'vscode-chat',
+}
+
+export type CopilotTokenResponse = {
+  token: string
+  expires_at: number
+  refresh_in: number
+  endpoints: {
+    api: string
+  }
+}
 
 export class GitHubDeviceFlowError extends Error {
   constructor(message: string) {
@@ -30,6 +46,8 @@ export type DeviceCodeResult = {
   interval: number
 }
 
+type FetchLike = (input: RequestInfo | URL, init?: RequestInit) => Promise<Response>
+
 export function getGithubDeviceFlowClientId(): string {
   return (
     process.env.GITHUB_DEVICE_FLOW_CLIENT_ID?.trim() ||
@@ -44,21 +62,21 @@ function sleep(ms: number): Promise<void> {
 export async function requestDeviceCode(options?: {
   clientId?: string
   scope?: string
-  fetchImpl?: typeof fetch
+  fetchImpl?: FetchLike
 }): Promise<DeviceCodeResult> {
   const clientId = options?.clientId ?? getGithubDeviceFlowClientId()
   if (!clientId) {
     throw new GitHubDeviceFlowError(
-      'No OAuth client ID: set GITHUB_DEVICE_FLOW_CLIENT_ID or paste a PAT instead.',
+      'No OAuth client ID: set GITHUB_DEVICE_FLOW_CLIENT_ID.',
     )
   }
   const fetchFn = options?.fetchImpl ?? fetch
   const requestedScope =
     options?.scope?.trim() || DEFAULT_GITHUB_DEVICE_SCOPE
   const scopesToTry =
-    requestedScope === OAUTH_SAFE_GITHUB_DEVICE_SCOPE
+    requestedScope === DEFAULT_GITHUB_DEVICE_SCOPE
       ? [requestedScope]
-      : [requestedScope, OAUTH_SAFE_GITHUB_DEVICE_SCOPE]
+      : [requestedScope, DEFAULT_GITHUB_DEVICE_SCOPE]
 
   let lastError = 'Device code request failed.'
 
@@ -77,7 +95,7 @@ export async function requestDeviceCode(options?: {
       lastError = `Device code request failed: ${res.status} ${text}`
       const isInvalidScope = /invalid_scope/i.test(text)
       const canRetryWithFallback =
-        scope !== OAUTH_SAFE_GITHUB_DEVICE_SCOPE && isInvalidScope
+        scope !== DEFAULT_GITHUB_DEVICE_SCOPE && isInvalidScope
       if (canRetryWithFallback) {
         continue
       }
@@ -114,7 +132,7 @@ export type PollOptions = {
   clientId?: string
   initialInterval?: number
   timeoutSeconds?: number
-  fetchImpl?: typeof fetch
+  fetchImpl?: FetchLike
 }
 
 export async function pollAccessToken(
@@ -197,3 +215,49 @@ export async function openVerificationUri(uri: string): Promise<void> {
     // User can open the URL manually
   }
 }
+
+/**
+ * Exchange an OAuth access token for a Copilot API token.
+ * The OAuth token alone cannot be used with the Copilot API endpoint.
+ */
+export async function exchangeForCopilotToken(
+  oauthToken: string,
+  fetchImpl?: FetchLike,
+): Promise<CopilotTokenResponse> {
+  const fetchFn = fetchImpl ?? fetch
+  const res = await fetchFn(COPILOT_TOKEN_URL, {
+    method: 'GET',
+    headers: {
+      Accept: 'application/json',
+      Authorization: `Bearer ${oauthToken}`,
+      ...COPILOT_HEADERS,
+    },
+  })
+  if (!res.ok) {
+    const text = await res.text().catch(() => '')
+    throw new GitHubDeviceFlowError(
+      `Copilot token exchange failed: ${res.status} ${text}`,
+    )
+  }
+  const data = (await res.json()) as Record<string, unknown>
+  const token = data.token
+  const expires_at = data.expires_at
+  const refresh_in = data.refresh_in
+  const endpoints = data.endpoints
+  if (
+    typeof token !== 'string' ||
+    typeof expires_at !== 'number' ||
+    typeof refresh_in !== 'number' ||
+    !endpoints ||
+    typeof endpoints !== 'object' ||
+    typeof (endpoints as Record<string, unknown>).api !== 'string'
+  ) {
+    throw new GitHubDeviceFlowError('Malformed Copilot token response')
+  }
+  return {
+    token,
+    expires_at,
+    refresh_in,
+    endpoints: endpoints as { api: string },
+  }
+}

src/services/mockRateLimits.ts

@@ -1,6 +1,11 @@
 // Mock rate limits for testing [internal-only]
 // The external build keeps this module as a stable no-op surface so imports
 // remain valid without exposing internal-only rate-limit simulation behavior.
+// This allows testing various rate limit scenarios without hitting actual limits
+//
+// WARNING: This is for internal testing/demo purposes only!
+// The mock headers may not exactly match the API specification or real-world behavior.
+// Always validate against actual API responses before relying on this for production features.
 
 import { setMockBillingAccessOverride } from '../utils/billing.js'
 import type { OverageDisabledReason } from './claudeAiLimits.js'

src/services/tips/tipRegistry.ts

@@ -645,7 +645,7 @@ const internalOnlyTips: Tip[] =
         {
           id: 'skillify',
           content: async () =>
-            '[internal] Turn repeatable workflows into reusable project skills when they keep recurring',
+            '[internal] Use /skillify to turn repeatable recurring workflows into reusable project skills',
           cooldownSessions: 15,
           isRelevant: async () => true,
         },

src/services/tools/toolExecution.test.ts

@@ -0,0 +1,33 @@
+import { describe, expect, test } from 'bun:test'
+
+import { SkillTool } from '../../tools/SkillTool/SkillTool.js'
+import {
+  getSchemaValidationErrorOverride,
+  getSchemaValidationToolUseResult,
+} from './toolExecution.js'
+
+describe('getSchemaValidationErrorOverride', () => {
+  test('returns actionable missing-skill error for SkillTool', () => {
+    expect(getSchemaValidationErrorOverride(SkillTool, {})).toBe(
+      'Missing skill name. Pass the slash command name as the skill parameter (e.g., skill: "commit" for /commit, skill: "review-pr" for /review-pr).',
+    )
+  })
+
+  test('does not override unrelated tool schema failures', () => {
+    expect(getSchemaValidationErrorOverride({ name: 'Read' } as never, {})).toBe(
+      null,
+    )
+  })
+
+  test('does not override SkillTool when skill is present', () => {
+    expect(
+      getSchemaValidationErrorOverride(SkillTool, { skill: 'commit' }),
+    ).toBe(null)
+  })
+
+  test('uses the actionable override for structured toolUseResult too', () => {
+    expect(getSchemaValidationToolUseResult(SkillTool, {} as never)).toBe(
+      'InputValidationError: Missing skill name. Pass the slash command name as the skill parameter (e.g., skill: "commit" for /commit, skill: "review-pr" for /review-pr).',
+    )
+  })
+})

src/services/tools/toolExecution.ts

@@ -43,6 +43,7 @@ import { FILE_READ_TOOL_NAME } from '../../tools/FileReadTool/prompt.js'
 import { FILE_WRITE_TOOL_NAME } from '../../tools/FileWriteTool/prompt.js'
 import { NOTEBOOK_EDIT_TOOL_NAME } from '../../tools/NotebookEditTool/constants.js'
 import { POWERSHELL_TOOL_NAME } from '../../tools/PowerShellTool/toolName.js'
+import { SKILL_TOOL_NAME } from '../../tools/SkillTool/constants.js'
 import { parseGitCommitId } from '../../tools/shared/gitOperationTracking.js'
 import {
   isDeferredTool,
@@ -596,6 +597,31 @@ export function buildSchemaNotSentHint(
   )
 }
 
+export function getSchemaValidationErrorOverride(
+  tool: Tool,
+  input: unknown,
+): string | null {
+  if (tool.name !== SKILL_TOOL_NAME || !input || typeof input !== 'object') {
+    return null
+  }
+
+  const skill = (input as { skill?: unknown }).skill
+  if (skill === undefined || skill === null) {
+    return 'Missing skill name. Pass the slash command name as the skill parameter (e.g., skill: "commit" for /commit, skill: "review-pr" for /review-pr).'
+  }
+
+  return null
+}
+
+export function getSchemaValidationToolUseResult(
+  tool: Tool,
+  input: unknown,
+  fallbackMessage?: string,
+): string {
+  const override = getSchemaValidationErrorOverride(tool, input)
+  return `InputValidationError: ${override ?? fallbackMessage ?? ''}`
+}
+
 async function checkPermissionsAndCallTool(
   tool: Tool,
   toolUseID: string,
@@ -614,7 +640,9 @@ async function checkPermissionsAndCallTool(
   // Validate input types with zod (surprisingly, the model is not great at generating valid input)
   const parsedInput = tool.inputSchema.safeParse(input)
   if (!parsedInput.success) {
-    let errorContent = formatZodValidationError(tool.name, parsedInput.error)
+    const fallbackErrorContent = formatZodValidationError(tool.name, parsedInput.error)
+    let errorContent =
+      getSchemaValidationErrorOverride(tool, input) ?? fallbackErrorContent
 
     const schemaHint = buildSchemaNotSentHint(
       tool,
@@ -672,7 +700,11 @@ async function checkPermissionsAndCallTool(
               tool_use_id: toolUseID,
             },
           ],
-          toolUseResult: `InputValidationError: ${parsedInput.error.message}`,
+          toolUseResult: getSchemaValidationToolUseResult(
+            tool,
+            input,
+            parsedInput.error.message,
+          ),
           sourceToolAssistantUUID: assistantMessage.uuid,
         }),
       },

src/services/tools/toolHooks.ts

@@ -29,6 +29,13 @@ import {
 } from '../../utils/permissions/PermissionResult.js'
 import { checkRuleBasedPermissions } from '../../utils/permissions/permissions.js'
 import { formatError } from '../../utils/toolErrors.js'
+import { getAutoFixConfig } from '../autoFix/autoFixConfig.js'
+import { shouldRunAutoFix, buildAutoFixContext } from '../autoFix/autoFixHook.js'
+import { runAutoFixCheck } from '../autoFix/autoFixRunner.js'
+
+// Track auto-fix retry count per query chain to enforce maxRetries cap.
+// Key: queryChainId (or 'default'), Value: number of auto-fix attempts used.
+const autoFixRetryCount = new Map<string, number>()
 import { isMcpTool } from '../mcp/utils.js'
 import type { McpServerType, MessageUpdateLazy } from './toolExecution.js'
 
@@ -185,6 +192,65 @@ export async function* runPostToolUseHooks<Input extends AnyObject, Output>(
         }
       }
     }
+
+    // Auto-fix: run lint/test if configured for this tool
+    const autoFixSettings = toolUseContext.getAppState().settings
+    const autoFixConfig = getAutoFixConfig(
+      autoFixSettings && typeof autoFixSettings === 'object' && 'autoFix' in autoFixSettings
+        ? (autoFixSettings as Record<string, unknown>).autoFix
+        : undefined,
+    )
+    if (shouldRunAutoFix(tool.name, autoFixConfig) && autoFixConfig) {
+      // Enforce maxRetries cap to prevent unbounded auto-fix loops.
+      // Uses queryChainId to scope the counter to the current conversation turn.
+      const chainKey = (toolUseContext.queryTracking?.chainId as string) ?? 'default'
+      const currentRetries = autoFixRetryCount.get(chainKey) ?? 0
+
+      if (currentRetries >= autoFixConfig.maxRetries) {
+        // Max retries reached — skip auto-fix and let the user know
+        yield {
+          message: createAttachmentMessage({
+            type: 'hook_additional_context',
+            content: [
+              `<auto_fix_feedback>\nAUTO-FIX: Maximum retry limit (${autoFixConfig.maxRetries}) reached. ` +
+              `Skipping further auto-fix attempts. Please review the errors manually.\n</auto_fix_feedback>`,
+            ],
+            hookName: `AutoFix:${tool.name}`,
+            toolUseID,
+            hookEvent: 'PostToolUse',
+          }),
+        }
+      } else {
+        try {
+          const cwd = toolUseContext.options?.cwd ?? process.cwd()
+          const autoFixResult = await runAutoFixCheck({
+            lint: autoFixConfig.lint,
+            test: autoFixConfig.test,
+            timeout: autoFixConfig.timeout,
+            cwd,
+            signal: toolUseContext.abortController.signal,
+          })
+          const autoFixContext = buildAutoFixContext(autoFixResult)
+          if (autoFixContext) {
+            autoFixRetryCount.set(chainKey, currentRetries + 1)
+            yield {
+              message: createAttachmentMessage({
+                type: 'hook_additional_context',
+                content: [autoFixContext],
+                hookName: `AutoFix:${tool.name}`,
+                toolUseID,
+                hookEvent: 'PostToolUse',
+              }),
+            }
+          } else {
+            // Lint/test passed — reset the retry counter for this chain
+            autoFixRetryCount.delete(chainKey)
+          }
+        } catch (autoFixError) {
+          logError(autoFixError)
+        }
+      }
+    }
   } catch (error) {
     logError(error)
   }

src/tools/AgentTool/prompt.ts

@@ -156,34 +156,24 @@ ${AGENT_TOOL_NAME}({
   const currentExamples = `Example usage:
 
 <example_agent_descriptions>
-"test-runner": use this agent after you are done writing code to run tests
+"claude-code-guide": use this agent when the user asks how Claude Code works or how to use its features
-"greeting-responder": use this agent to respond to user greetings with a friendly joke
+"statusline-setup": use this agent to configure the user's Claude Code status line setting
 </example_agent_descriptions>
 
 <example>
-user: "Please write a function that checks if a number is prime"
+user: "How do I configure Claude Code hooks?"
-assistant: I'm going to use the ${FILE_WRITE_TOOL_NAME} tool to write the following code:
-<code>
-function isPrime(n) {
-  if (n <= 1) return false
-  for (let i = 2; i * i <= n; i++) {
-    if (n % i === 0) return false
-  }
-  return true
-}
-</code>
 <commentary>
-Since a significant piece of code was written and the task was completed, now use the test-runner agent to run the tests
+This is a Claude Code usage question, so use the claude-code-guide agent
 </commentary>
-assistant: Uses the ${AGENT_TOOL_NAME} tool to launch the test-runner agent
+assistant: Uses the ${AGENT_TOOL_NAME} tool to launch the claude-code-guide agent
 </example>
 
 <example>
-user: "Hello"
+user: "Set up my Claude Code status line"
 <commentary>
-Since the user is greeting, use the greeting-responder agent to respond with a friendly joke
+This matches the statusline-setup agent, so use it to configure the setting
 </commentary>
-assistant: "I'm going to use the ${AGENT_TOOL_NAME} tool to launch the greeting-responder agent"
+assistant: "I'm going to use the ${AGENT_TOOL_NAME} tool to launch the statusline-setup agent"
 </example>
 `
 

src/tools/SkillTool/SkillTool.test.ts

@@ -0,0 +1,31 @@
+import { describe, expect, test } from 'bun:test'
+
+import { SkillTool } from './SkillTool.js'
+
+describe('SkillTool missing parameter handling', () => {
+  test('missing skill stays required at the schema level', async () => {
+    const parsed = SkillTool.inputSchema.safeParse({})
+
+    expect(parsed.success).toBe(false)
+  })
+
+  test('validateInput still returns an actionable error when called with missing skill', async () => {
+    const result = await SkillTool.validateInput?.({} as never, {
+      options: { tools: [] },
+      messages: [],
+    } as never)
+
+    expect(result).toEqual({
+      result: false,
+      message:
+        'Missing skill name. Pass the slash command name as the skill parameter (e.g., skill: "commit" for /commit, skill: "review-pr" for /review-pr).',
+      errorCode: 1,
+    })
+  })
+
+  test('valid skill input still parses and validates', async () => {
+    const parsed = SkillTool.inputSchema.safeParse({ skill: 'commit' })
+
+    expect(parsed.success).toBe(true)
+  })
+})

src/tools/SkillTool/SkillTool.ts

@@ -352,6 +352,16 @@ export const SkillTool: Tool<InputSchema, Output, Progress> = buildTool({
   toAutoClassifierInput: ({ skill }) => skill ?? '',
 
   async validateInput({ skill }, context): Promise<ValidationResult> {
+    if (!skill || typeof skill !== 'string') {
+      return {
+        result: false,
+        message:
+          'Missing skill name. Pass the slash command name as the skill parameter ' +
+          '(e.g., skill: "commit" for /commit, skill: "review-pr" for /review-pr).',
+        errorCode: 1,
+      }
+    }
+
     // Skills are just skill names, no arguments
     const trimmed = skill.trim()
     if (!trimmed) {
@@ -434,7 +444,7 @@ export const SkillTool: Tool<InputSchema, Output, Progress> = buildTool({
     context,
   ): Promise<PermissionDecision> {
     // Skills are just skill names, no arguments
-    const trimmed = skill.trim()
+    const trimmed = skill ?? ''
 
     // Remove leading slash if present (for compatibility)
     const commandName = trimmed.startsWith('/') ? trimmed.substring(1) : trimmed
@@ -592,7 +602,7 @@ export const SkillTool: Tool<InputSchema, Output, Progress> = buildTool({
     // - Skill is a prompt-based skill
 
     // Skills are just names, with optional arguments
-    const trimmed = skill.trim()
+    const trimmed = skill ?? ''
 
     // Remove leading slash if present (for compatibility)
     const commandName = trimmed.startsWith('/') ? trimmed.substring(1) : trimmed

src/utils/api.test.ts

@@ -1,6 +1,7 @@
 import { expect, test } from 'bun:test'
 import { z } from 'zod/v4'
 import { getEmptyToolPermissionContext, type Tool, type Tools } from '../Tool.js'
+import { SkillTool } from '../tools/SkillTool/SkillTool.js'
 import { toolToAPISchema } from './api.js'
 
 test('toolToAPISchema preserves provider-specific schema keywords in input_schema', async () => {
@@ -64,3 +65,16 @@ test('toolToAPISchema preserves provider-specific schema keywords in input_schem
     },
   })
 })
+
+test('toolToAPISchema keeps skill required for SkillTool', async () => {
+  const schema = await toolToAPISchema(SkillTool, {
+    getToolPermissionContext: async () => getEmptyToolPermissionContext(),
+    tools: [] as unknown as Tools,
+    agents: [],
+  })
+
+  expect((schema as { input_schema: unknown }).input_schema).toMatchObject({
+    type: 'object',
+    required: ['skill'],
+  })
+})

src/utils/context.test.ts

@@ -94,3 +94,22 @@ test('gpt-5.4 family keeps large max output overrides within provider limits', (
   expect(getMaxOutputTokensForModel('gpt-5.4-mini')).toBe(128_000)
   expect(getMaxOutputTokensForModel('gpt-5.4-nano')).toBe(128_000)
 })
+
+test('MiniMax-M2.7 uses explicit provider-specific context and output caps', () => {
+  process.env.CLAUDE_CODE_USE_OPENAI = '1'
+  delete process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS
+
+  expect(getContextWindowForModel('MiniMax-M2.7')).toBe(204_800)
+  expect(getModelMaxOutputTokens('MiniMax-M2.7')).toEqual({
+    default: 131_072,
+    upperLimit: 131_072,
+  })
+  expect(getMaxOutputTokensForModel('MiniMax-M2.7')).toBe(131_072)
+})
+
+test('unknown openai-compatible models still use the conservative fallback window', () => {
+  process.env.CLAUDE_CODE_USE_OPENAI = '1'
+  delete process.env.CLAUDE_CODE_MAX_OUTPUT_TOKENS
+
+  expect(getContextWindowForModel('some-unknown-3p-model')).toBe(8_000)
+})

src/utils/context.ts

@@ -72,16 +72,23 @@ export function getContextWindowForModel(
     return 1_000_000
   }
 
-  // OpenAI-compatible provider — use known context windows for the model
+  // OpenAI-compatible provider — use known context windows for the model.
-  if (
+  // Unknown models get a conservative 8k default so auto-compact triggers
+  // before hitting a hard context_window_exceeded error.
+  const isOpenAIProvider =
     isEnvTruthy(process.env.CLAUDE_CODE_USE_OPENAI) ||
     isEnvTruthy(process.env.CLAUDE_CODE_USE_GEMINI) ||
     isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB)
-  ) {
+  if (isOpenAIProvider) {
     const openaiWindow = getOpenAIContextWindow(model)
     if (openaiWindow !== undefined) {
       return openaiWindow
     }
+    console.error(
+      `[context] Warning: model "${model}" not in context window table — using conservative 8k default. ` +
+      'Add it to src/utils/model/openaiContextWindows.ts for accurate compaction.',
+    )
+    return 8_000
   }
 
   const cap = getModelCapability(model)

src/utils/conversationRecovery.hooks.test.ts

@@ -69,3 +69,93 @@ test('loadConversationForResume rejects oversized transcripts before resume hook
   )
   expect(hookSpy).not.toHaveBeenCalled()
 })
+
+test('deserializeMessagesWithInterruptDetection strips thinking blocks only for OpenAI-compatible providers', async () => {
+  const serializedMessages = [
+    user(id(10), 'hello'),
+    {
+      type: 'assistant',
+      uuid: id(11),
+      parentUuid: id(10),
+      timestamp: ts,
+      cwd: '/tmp',
+      sessionId,
+      version: 'test',
+      message: {
+        role: 'assistant',
+        content: [
+          { type: 'thinking', thinking: 'secret reasoning' },
+          { type: 'text', text: 'visible reply' },
+        ],
+      },
+    },
+    {
+      type: 'assistant',
+      uuid: id(12),
+      parentUuid: id(11),
+      timestamp: ts,
+      cwd: '/tmp',
+      sessionId,
+      version: 'test',
+      message: {
+        role: 'assistant',
+        content: [{ type: 'thinking', thinking: 'only hidden reasoning' }],
+      },
+    },
+    user(id(13), 'follow up'),
+  ]
+
+  mock.module('./model/providers.js', () => ({
+    getAPIProvider: () => 'openai',
+    isOpenAICompatibleProvider: (provider: string) =>
+      provider === 'openai' ||
+      provider === 'gemini' ||
+      provider === 'github' ||
+      provider === 'codex',
+  }))
+
+  const openaiModule = await import(`./conversationRecovery.ts?provider=openai-${Date.now()}`)
+  const thirdParty = openaiModule.deserializeMessagesWithInterruptDetection(serializedMessages as never[])
+  const thirdPartyAssistantMessages = thirdParty.messages.filter(
+    message => message.type === 'assistant',
+  )
+
+  expect(thirdPartyAssistantMessages).toHaveLength(2)
+  expect(thirdPartyAssistantMessages[0]?.message?.content).toEqual([
+    { type: 'text', text: 'visible reply' },
+  ])
+  expect(
+    JSON.stringify(thirdPartyAssistantMessages.map(message => message.message?.content)),
+  ).not.toContain('secret reasoning')
+  expect(
+    JSON.stringify(thirdPartyAssistantMessages.map(message => message.message?.content)),
+  ).not.toContain('only hidden reasoning')
+
+  mock.restore()
+  mock.module('./model/providers.js', () => ({
+    getAPIProvider: () => 'bedrock',
+    isOpenAICompatibleProvider: (provider: string) =>
+      provider === 'openai' ||
+      provider === 'gemini' ||
+      provider === 'github' ||
+      provider === 'codex',
+  }))
+
+  const bedrockModule = await import(`./conversationRecovery.ts?provider=bedrock-${Date.now()}`)
+  const anthropicCompatible = bedrockModule.deserializeMessagesWithInterruptDetection(serializedMessages as never[])
+  const anthropicAssistantMessages = anthropicCompatible.messages.filter(
+    message => message.type === 'assistant',
+  )
+
+  expect(anthropicAssistantMessages).toHaveLength(2)
+  expect(anthropicAssistantMessages[0]?.message?.content).toEqual([
+    { type: 'thinking', thinking: 'secret reasoning' },
+    { type: 'text', text: 'visible reply' },
+  ])
+  expect(
+    JSON.stringify(anthropicAssistantMessages.map(message => message.message?.content)),
+  ).toContain('secret reasoning')
+  expect(
+    JSON.stringify(anthropicAssistantMessages.map(message => message.message?.content)),
+  ).not.toContain('only hidden reasoning')
+})

src/utils/conversationRecovery.test.ts

@@ -13,6 +13,7 @@ const originalSimple = process.env.CLAUDE_CODE_SIMPLE
 const sessionId = '00000000-0000-4000-8000-000000001999'
 const ts = '2026-04-02T00:00:00.000Z'
 
+
 function id(n: number): string {
   return `00000000-0000-4000-8000-${String(n).padStart(12, '0')}`
 }
@@ -76,4 +77,3 @@ test('loadConversationForResume rejects oversized reconstructed transcripts', as
     'Reconstructed transcript is too large to resume safely',
   )
 })
-

src/utils/conversationRecovery.ts

@@ -24,6 +24,7 @@ import {
   type FileHistorySnapshot,
 } from './fileHistory.js'
 import { logError } from './log.js'
+import { getAPIProvider } from './model/providers.js'
 import {
   createAssistantMessage,
   createUserMessage,
@@ -177,6 +178,25 @@ export type DeserializeResult = {
   turnInterruptionState: TurnInterruptionState
 }
 
+/**
+ * Remove thinking/redacted_thinking content blocks from assistant messages.
+ * Messages that become empty after stripping are removed entirely.
+ */
+function stripThinkingBlocks(messages: NormalizedMessage[]): NormalizedMessage[] {
+  return messages.reduce<NormalizedMessage[]>((acc, msg) => {
+    if (msg.type !== 'assistant' || !Array.isArray(msg.message?.content)) {
+      acc.push(msg)
+      return acc
+    }
+    const filtered = msg.message.content.filter(
+      (block: { type?: string }) => block.type !== 'thinking' && block.type !== 'redacted_thinking',
+    )
+    if (filtered.length === 0) return acc
+    acc.push({ ...msg, message: { ...msg.message, content: filtered } })
+    return acc
+  }, [])
+}
+
 /**
  * Deserializes messages from a log file into the format expected by the REPL.
  * Filters unresolved tool uses, orphaned thinking messages, and appends a
@@ -227,10 +247,19 @@ export function deserializeMessagesWithInterruptDetection(
       filteredToolUses,
     ) as NormalizedMessage[]
 
+    // Strip thinking/redacted_thinking content blocks from assistant messages
+    // when resuming against a 3P provider. These Anthropic-specific blocks cause
+    // 400 errors or context corruption on OpenAI-compatible providers (issue #248 finding 5).
+    const provider = getAPIProvider()
+    const isThirdPartyProvider = provider !== 'firstParty' && provider !== 'bedrock' && provider !== 'vertex' && provider !== 'foundry'
+    const thinkingStripped = isThirdPartyProvider
+      ? stripThinkingBlocks(filteredThinking)
+      : filteredThinking
+
     // Filter out assistant messages with only whitespace text content.
     // This can happen when model outputs "\n\n" before thinking, user cancels mid-stream.
     const filteredMessages = filterWhitespaceOnlyAssistantMessages(
-      filteredThinking,
+      thinkingStripped,
     ) as NormalizedMessage[]
 
     const internalState = detectTurnInterruption(filteredMessages)

src/utils/dragDropPaths.test.ts

@@ -4,6 +4,10 @@ import { tmpdir } from 'os'
 import { join } from 'path'
 import { extractDraggedFilePaths } from './dragDropPaths.js'
 
+function escapeFinderDraggedPath(filePath: string): string {
+  return filePath.replace(/([\\ ])/g, '\\$1')
+}
+
 describe('extractDraggedFilePaths', () => {
   // Paths that exist on any system.
   const thisFile = import.meta.path
@@ -80,6 +84,12 @@ describe('extractDraggedFilePaths', () => {
     })
   })
 
+  test('escapeFinderDraggedPath escapes spaces and backslashes', () => {
+    expect(escapeFinderDraggedPath('/tmp/my\\notes file.txt')).toBe(
+      '/tmp/my\\\\notes\\ file.txt',
+    )
+  })
+
   // Backslash-escaped paths are a Finder/macOS + Linux convention — on
   // Windows the shell-escape step is skipped, so these cases do not apply.
   if (process.platform !== 'win32') {
@@ -92,7 +102,7 @@ describe('extractDraggedFilePaths', () => {
 
       test('resolves an escaped real file with a space in its name', () => {
         // Raw form matches what a terminal delivers on Finder drag.
-        const escaped = spacedFile.replace(/ /g, '\\ ')
+        const escaped = escapeFinderDraggedPath(spacedFile)
         expect(extractDraggedFilePaths(escaped)).toEqual([spacedFile])
       })
     })

src/utils/geminiCredentials.test.ts

@@ -3,6 +3,7 @@ import { afterEach, beforeEach, expect, mock, test } from 'bun:test'
 type MockStorageData = Record<string, unknown>
 
 const originalEnv = { ...process.env }
+const originalArgv = [...process.argv]
 let storageState: MockStorageData = {}
 
 async function importFreshModule() {
@@ -27,11 +28,14 @@ async function importFreshModule() {
 
 beforeEach(() => {
   process.env = { ...originalEnv }
+  delete process.env.CLAUDE_CODE_SIMPLE
+  process.argv = originalArgv.filter(arg => arg !== '--bare')
   storageState = {}
 })
 
 afterEach(() => {
   process.env = { ...originalEnv }
+  process.argv = [...originalArgv]
   storageState = {}
   mock.restore()
 })

src/utils/githubModelsCredentials.hydrate.test.ts

@@ -41,7 +41,7 @@ describe('hydrateGithubModelsTokenFromSecureStorage', () => {
     }))
 
     const { hydrateGithubModelsTokenFromSecureStorage } = await import(
-      './githubModelsCredentials.js'
+      './githubModelsCredentials.js?hydrate=sets-token'
     )
     hydrateGithubModelsTokenFromSecureStorage()
     expect(process.env.GITHUB_TOKEN).toBe('stored-secret')
@@ -62,7 +62,7 @@ describe('hydrateGithubModelsTokenFromSecureStorage', () => {
     }))
 
     const { hydrateGithubModelsTokenFromSecureStorage } = await import(
-      './githubModelsCredentials.js'
+      './githubModelsCredentials.js?hydrate=preserve-existing'
     )
     hydrateGithubModelsTokenFromSecureStorage()
     expect(process.env.GITHUB_TOKEN).toBe('already')

src/utils/githubModelsCredentials.refresh.test.ts

@@ -0,0 +1,118 @@
+import { afterEach, beforeEach, describe, expect, mock, test } from 'bun:test'
+
+async function importFreshModule() {
+  mock.restore()
+  return import(`./githubModelsCredentials.ts?ts=${Date.now()}-${Math.random()}`)
+}
+
+describe('refreshGithubModelsTokenIfNeeded', () => {
+  const orig = {
+    CLAUDE_CODE_USE_GITHUB: process.env.CLAUDE_CODE_USE_GITHUB,
+    CLAUDE_CODE_SIMPLE: process.env.CLAUDE_CODE_SIMPLE,
+    GITHUB_TOKEN: process.env.GITHUB_TOKEN,
+    GH_TOKEN: process.env.GH_TOKEN,
+  }
+
+  beforeEach(() => {
+    mock.restore()
+  })
+
+  afterEach(() => {
+    for (const [k, v] of Object.entries(orig)) {
+      if (v === undefined) {
+        delete process.env[k as keyof typeof orig]
+      } else {
+        process.env[k as keyof typeof orig] = v
+      }
+    }
+  })
+
+  test('refreshes expired Copilot token using stored OAuth token', async () => {
+    process.env.CLAUDE_CODE_USE_GITHUB = '1'
+    delete process.env.CLAUDE_CODE_SIMPLE
+    delete process.env.GITHUB_TOKEN
+    delete process.env.GH_TOKEN
+
+    const futureExp = Math.floor(Date.now() / 1000) + 3600
+    let store: Record<string, unknown> = {
+      githubModels: {
+        accessToken: 'tid=stale;exp=1;sku=free',
+        oauthAccessToken: 'ghu_oauth_secret',
+      },
+    }
+
+    mock.module('./secureStorage/index.js', () => ({
+      getSecureStorage: () => ({
+        read: () => store,
+        update: (next: Record<string, unknown>) => {
+          store = next
+          return { success: true }
+        },
+      }),
+    }))
+
+    mock.module('../services/github/deviceFlow.js', () => ({
+      DEFAULT_GITHUB_DEVICE_SCOPE: 'read:user',
+      exchangeForCopilotToken: async () => ({
+        token: `tid=fresh;exp=${futureExp};sku=free`,
+        expires_at: futureExp,
+        refresh_in: 1500,
+        endpoints: { api: 'https://api.githubcopilot.com' },
+      }),
+    }))
+
+    const { refreshGithubModelsTokenIfNeeded } = await importFreshModule()
+
+    const refreshed = await refreshGithubModelsTokenIfNeeded()
+    expect(refreshed).toBe(true)
+    expect(process.env.GITHUB_TOKEN?.startsWith('tid=fresh;exp=')).toBe(true)
+
+    const githubModels = (store.githubModels ?? {}) as {
+      accessToken?: string
+      oauthAccessToken?: string
+    }
+    expect(githubModels.accessToken?.startsWith('tid=fresh;exp=')).toBe(true)
+    expect(githubModels.oauthAccessToken).toBe('ghu_oauth_secret')
+  })
+
+  test('does not refresh when current Copilot token is valid', async () => {
+    process.env.CLAUDE_CODE_USE_GITHUB = '1'
+    delete process.env.CLAUDE_CODE_SIMPLE
+    delete process.env.GITHUB_TOKEN
+    delete process.env.GH_TOKEN
+
+    const futureExp = Math.floor(Date.now() / 1000) + 3600
+    const exchangeSpy = mock(async () => ({
+      token: `tid=unexpected;exp=${futureExp};sku=free`,
+      expires_at: futureExp,
+      refresh_in: 1500,
+      endpoints: { api: 'https://api.githubcopilot.com' },
+    }))
+
+    mock.module('./secureStorage/index.js', () => ({
+      getSecureStorage: () => ({
+        read: () => ({
+          githubModels: {
+            accessToken: `tid=already-valid;exp=${futureExp};sku=free`,
+            oauthAccessToken: 'ghu_oauth_secret',
+          },
+        }),
+        update: () => ({ success: true }),
+      }),
+    }))
+
+    mock.module('../services/github/deviceFlow.js', () => ({
+      DEFAULT_GITHUB_DEVICE_SCOPE: 'read:user',
+      exchangeForCopilotToken: exchangeSpy,
+    }))
+
+    const { refreshGithubModelsTokenIfNeeded } = await importFreshModule()
+
+    const refreshed = await refreshGithubModelsTokenIfNeeded()
+    expect(refreshed).toBe(false)
+    expect(exchangeSpy).not.toHaveBeenCalled()
+    expect(process.env.GITHUB_TOKEN?.startsWith('tid=already-valid;exp=')).toBe(
+      true,
+    )
+  })
+})

src/utils/githubModelsCredentials.test.ts

@@ -1,13 +1,11 @@
 import { describe, expect, test } from 'bun:test'
 
-import {
-  clearGithubModelsToken,
-  readGithubModelsToken,
-  saveGithubModelsToken,
-} from './githubModelsCredentials.js'
-
 describe('readGithubModelsToken', () => {
-  test('returns undefined in bare mode', () => {
+  test('returns undefined in bare mode', async () => {
+    const { readGithubModelsToken } = await import(
+      './githubModelsCredentials.js?read-bare-mode'
+    )
+
     const prev = process.env.CLAUDE_CODE_SIMPLE
     process.env.CLAUDE_CODE_SIMPLE = '1'
     expect(readGithubModelsToken()).toBeUndefined()
@@ -20,7 +18,11 @@ describe('readGithubModelsToken', () => {
 })
 
 describe('saveGithubModelsToken / clearGithubModelsToken', () => {
-  test('save returns failure in bare mode', () => {
+  test('save returns failure in bare mode', async () => {
+    const { saveGithubModelsToken } = await import(
+      './githubModelsCredentials.js?save-bare-mode'
+    )
+
     const prev = process.env.CLAUDE_CODE_SIMPLE
     process.env.CLAUDE_CODE_SIMPLE = '1'
     const r = saveGithubModelsToken('abc')
@@ -33,7 +35,11 @@ describe('saveGithubModelsToken / clearGithubModelsToken', () => {
     }
   })
 
-  test('clear succeeds in bare mode', () => {
+  test('clear succeeds in bare mode', async () => {
+    const { clearGithubModelsToken } = await import(
+      './githubModelsCredentials.js?clear-bare-mode'
+    )
+
     const prev = process.env.CLAUDE_CODE_SIMPLE
     process.env.CLAUDE_CODE_SIMPLE = '1'
     expect(clearGithubModelsToken().success).toBe(true)

src/utils/githubModelsCredentials.ts

@@ -1,5 +1,6 @@
 import { isBareMode, isEnvTruthy } from './envUtils.js'
 import { getSecureStorage } from './secureStorage/index.js'
+import { exchangeForCopilotToken } from '../services/github/deviceFlow.js'
 
 /** JSON key in the shared OpenClaude secure storage blob. */
 export const GITHUB_MODELS_STORAGE_KEY = 'githubModels' as const
@@ -8,6 +9,38 @@ export const GITHUB_MODELS_HYDRATED_ENV_MARKER =
 
 export type GithubModelsCredentialBlob = {
   accessToken: string
+  oauthAccessToken?: string
+}
+
+type GithubTokenStatus = 'valid' | 'expired' | 'invalid_format'
+
+function checkGithubTokenStatus(token: string): GithubTokenStatus {
+  const expMatch = token.match(/exp=(\d+)/)
+  if (expMatch) {
+    const expSeconds = Number(expMatch[1])
+    if (!Number.isNaN(expSeconds)) {
+      return Date.now() >= expSeconds * 1000 ? 'expired' : 'valid'
+    }
+  }
+
+  const parts = token.split('.')
+  const looksLikeJwt =
+    parts.length === 3 && parts.every(part => /^[A-Za-z0-9_-]+$/.test(part))
+  if (looksLikeJwt) {
+    try {
+      const normalized = parts[1].replace(/-/g, '+').replace(/_/g, '/')
+      const padded = normalized + '='.repeat((4 - (normalized.length % 4)) % 4)
+      const json = Buffer.from(padded, 'base64').toString('utf8')
+      const parsed = JSON.parse(json)
+      if (parsed && typeof parsed === 'object' && parsed.exp) {
+        return Date.now() >= (parsed.exp as number) * 1000 ? 'expired' : 'valid'
+      }
+    } catch {
+      return 'invalid_format'
+    }
+  }
+
+  return 'invalid_format'
 }
 
 export function readGithubModelsToken(): string | undefined {
@@ -23,6 +56,19 @@ export function readGithubModelsToken(): string | undefined {
   }
 }
 
+export async function readGithubModelsTokenAsync(): Promise<string | undefined> {
+  if (isBareMode()) return undefined
+  try {
+    const data = (await getSecureStorage().readAsync()) as
+      | ({ githubModels?: GithubModelsCredentialBlob } & Record<string, unknown>)
+      | null
+    const t = data?.githubModels?.accessToken?.trim()
+    return t || undefined
+  } catch {
+    return undefined
+  }
+}
+
 /**
  * If GitHub Models mode is on and no token is in the environment, copy the
  * stored token into process.env so the OpenAI shim and validation see it.
@@ -53,7 +99,62 @@ export function hydrateGithubModelsTokenFromSecureStorage(): void {
   delete process.env[GITHUB_MODELS_HYDRATED_ENV_MARKER]
 }
 
-export function saveGithubModelsToken(token: string): {
+/**
+ * Startup auto-refresh for GitHub Models mode.
+ *
+ * If a stored Copilot token is expired/invalid and an OAuth token is present,
+ * exchange the OAuth token for a fresh Copilot token and persist it.
+ */
+export async function refreshGithubModelsTokenIfNeeded(): Promise<boolean> {
+  if (!isEnvTruthy(process.env.CLAUDE_CODE_USE_GITHUB)) {
+    return false
+  }
+  if (isBareMode()) {
+    return false
+  }
+
+  try {
+    const secureStorage = getSecureStorage()
+    const data = secureStorage.read() as
+      | ({ githubModels?: GithubModelsCredentialBlob } & Record<string, unknown>)
+      | null
+    const blob = data?.githubModels
+    const accessToken = blob?.accessToken?.trim() || ''
+    const oauthToken = blob?.oauthAccessToken?.trim() || ''
+
+    if (!accessToken && !oauthToken) {
+      return false
+    }
+
+    const status = accessToken ? checkGithubTokenStatus(accessToken) : 'expired'
+    if (status === 'valid') {
+      if (!process.env.GITHUB_TOKEN?.trim() && !process.env.GH_TOKEN?.trim()) {
+        process.env.GITHUB_TOKEN = accessToken
+      }
+      return false
+    }
+
+    if (!oauthToken) {
+      return false
+    }
+
+    const refreshed = await exchangeForCopilotToken(oauthToken)
+    const saved = saveGithubModelsToken(refreshed.token, oauthToken)
+    if (!saved.success) {
+      return false
+    }
+
+    process.env.GITHUB_TOKEN = refreshed.token
+    return true
+  } catch {
+    return false
+  }
+}
+
+export function saveGithubModelsToken(
+  token: string,
+  oauthToken?: string,
+): {
   success: boolean
   warning?: string
 } {
@@ -66,9 +167,21 @@ export function saveGithubModelsToken(token: string): {
   }
   const secureStorage = getSecureStorage()
   const prev = secureStorage.read() || {}
+  const prevGithubModels = (prev as Record<string, unknown>)[
+    GITHUB_MODELS_STORAGE_KEY
+  ] as GithubModelsCredentialBlob | undefined
+  const oauthTrimmed = oauthToken?.trim()
+  const mergedBlob: GithubModelsCredentialBlob = {
+    accessToken: trimmed,
+  }
+  if (oauthTrimmed) {
+    mergedBlob.oauthAccessToken = oauthTrimmed
+  } else if (prevGithubModels?.oauthAccessToken?.trim()) {
+    mergedBlob.oauthAccessToken = prevGithubModels.oauthAccessToken.trim()
+  }
   const merged = {
     ...(prev as Record<string, unknown>),
-    [GITHUB_MODELS_STORAGE_KEY]: { accessToken: trimmed },
+    [GITHUB_MODELS_STORAGE_KEY]: mergedBlob,
   }
   return secureStorage.update(merged as typeof prev)
 }

src/utils/model/configs.ts

@@ -35,6 +35,8 @@ export const CLAUDE_3_7_SONNET_CONFIG = {
   foundry: 'claude-3-7-sonnet',
   openai: 'gpt-4o-mini',
   gemini: 'gemini-2.0-flash',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_3_5_V2_SONNET_CONFIG = {
@@ -44,6 +46,8 @@ export const CLAUDE_3_5_V2_SONNET_CONFIG = {
   foundry: 'claude-3-5-sonnet',
   openai: 'gpt-4o-mini',
   gemini: 'gemini-2.0-flash',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_3_5_HAIKU_CONFIG = {
@@ -53,6 +57,8 @@ export const CLAUDE_3_5_HAIKU_CONFIG = {
   foundry: 'claude-3-5-haiku',
   openai: 'gpt-4o-mini',
   gemini: 'gemini-2.0-flash-lite',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_HAIKU_4_5_CONFIG = {
@@ -62,6 +68,8 @@ export const CLAUDE_HAIKU_4_5_CONFIG = {
   foundry: 'claude-haiku-4-5',
   openai: 'gpt-4o-mini',
   gemini: 'gemini-2.0-flash-lite',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_SONNET_4_CONFIG = {
@@ -71,6 +79,8 @@ export const CLAUDE_SONNET_4_CONFIG = {
   foundry: 'claude-sonnet-4',
   openai: 'gpt-4o-mini',
   gemini: 'gemini-2.0-flash',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_SONNET_4_5_CONFIG = {
@@ -80,6 +90,8 @@ export const CLAUDE_SONNET_4_5_CONFIG = {
   foundry: 'claude-sonnet-4-5',
   openai: 'gpt-4o',
   gemini: 'gemini-2.0-flash',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_OPUS_4_CONFIG = {
@@ -89,6 +101,8 @@ export const CLAUDE_OPUS_4_CONFIG = {
   foundry: 'claude-opus-4',
   openai: 'gpt-4o',
   gemini: 'gemini-2.5-pro-preview-03-25',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_OPUS_4_1_CONFIG = {
@@ -98,6 +112,8 @@ export const CLAUDE_OPUS_4_1_CONFIG = {
   foundry: 'claude-opus-4-1',
   openai: 'gpt-4o',
   gemini: 'gemini-2.5-pro-preview-03-25',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_OPUS_4_5_CONFIG = {
@@ -107,6 +123,8 @@ export const CLAUDE_OPUS_4_5_CONFIG = {
   foundry: 'claude-opus-4-5',
   openai: 'gpt-4o',
   gemini: 'gemini-2.5-pro-preview-03-25',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_OPUS_4_6_CONFIG = {
@@ -116,6 +134,8 @@ export const CLAUDE_OPUS_4_6_CONFIG = {
   foundry: 'claude-opus-4-6',
   openai: 'gpt-4o',
   gemini: 'gemini-2.5-pro-preview-03-25',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 export const CLAUDE_SONNET_4_6_CONFIG = {
@@ -125,6 +145,8 @@ export const CLAUDE_SONNET_4_6_CONFIG = {
   foundry: 'claude-sonnet-4-6',
   openai: 'gpt-4o',
   gemini: 'gemini-2.0-flash',
+  github: 'github:copilot',
+  codex: 'gpt-5.4',
 } as const satisfies ModelConfig
 
 // @[MODEL LAUNCH]: Register the new config here.

src/utils/model/copilotModels.ts

@@ -0,0 +1,351 @@
+/**
+ * Hardcoded Copilot model registry from models.dev/api.json
+ * These are the 19 models available through GitHub Copilot.
+ */
+
+export type CopilotModel = {
+  id: string
+  name: string
+  family: string
+  attachment: boolean
+  reasoning: boolean
+  tool_call: boolean
+  temperature: boolean
+  knowledge: string
+  release_date: string
+  last_updated: string
+  modalities: {
+    input: string[]
+    output: string[]
+  }
+  open_weights: boolean
+  cost: {
+    input: number
+    output: number
+    cache_read?: number
+  }
+  limit: {
+    context: number
+    input?: number
+    output: number
+  }
+}
+
+export const COPILOT_MODELS: Record<string, CopilotModel> = {
+  'gpt-5.4': {
+    id: 'gpt-5.4',
+    name: 'GPT-5.4',
+    family: 'gpt',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 400000, output: 32768 },
+  },
+  'gpt-5.4-mini': {
+    id: 'gpt-5.4-mini',
+    name: 'GPT-5.4 mini',
+    family: 'gpt-mini',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 400000, output: 32768 },
+  },
+  'gpt-5.3-codex': {
+    id: 'gpt-5.3-codex',
+    name: 'GPT-5.3-Codex',
+    family: 'gpt-codex',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 400000, output: 32768 },
+  },
+  'gpt-5.2-codex': {
+    id: 'gpt-5.2-codex',
+    name: 'GPT-5.2-Codex',
+    family: 'gpt-codex',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 400000, output: 32768 },
+  },
+  'gpt-5.2': {
+    id: 'gpt-5.2',
+    name: 'GPT-5.2',
+    family: 'gpt',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 264000, output: 32768 },
+  },
+  'gpt-5.1-codex': {
+    id: 'gpt-5.1-codex',
+    name: 'GPT-5.1-Codex',
+    family: 'gpt-codex',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 400000, output: 32768 },
+  },
+  'gpt-5.1-codex-max': {
+    id: 'gpt-5.1-codex-max',
+    name: 'GPT-5.1-Codex-max',
+    family: 'gpt-codex',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 400000, output: 32768 },
+  },
+  'gpt-5.1-codex-mini': {
+    id: 'gpt-5.1-codex-mini',
+    name: 'GPT-5.1-Codex-mini',
+    family: 'gpt-codex',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 400000, output: 32768 },
+  },
+  'gpt-4o': {
+    id: 'gpt-4o',
+    name: 'GPT-4o',
+    family: 'gpt',
+    attachment: true,
+    reasoning: false,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2023-10',
+    release_date: '2024-05-01',
+    last_updated: '2024-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 128000, output: 16384 },
+  },
+  'gpt-4.1': {
+    id: 'gpt-4.1',
+    name: 'GPT-4.1',
+    family: 'gpt',
+    attachment: false,
+    reasoning: false,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2024-06',
+    release_date: '2024-06-01',
+    last_updated: '2024-06-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 128000, output: 32768 },
+  },
+  'claude-opus-4.6': {
+    id: 'claude-opus-4.6',
+    name: 'Claude Opus 4.6',
+    family: 'claude-opus',
+    attachment: true,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 144000, output: 32768 },
+  },
+  'claude-opus-4.5': {
+    id: 'claude-opus-4.5',
+    name: 'Claude Opus 4.5',
+    family: 'claude-opus',
+    attachment: true,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 160000, output: 32768 },
+  },
+  'claude-sonnet-4.6': {
+    id: 'claude-sonnet-4.6',
+    name: 'Claude Sonnet 4.6',
+    family: 'claude-sonnet',
+    attachment: true,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 200000, output: 32768 },
+  },
+  'claude-sonnet-4.5': {
+    id: 'claude-sonnet-4.5',
+    name: 'Claude Sonnet 4.5',
+    family: 'claude-sonnet',
+    attachment: true,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 144000, output: 32768 },
+  },
+  'claude-haiku-4.5': {
+    id: 'claude-haiku-4.5',
+    name: 'Claude Haiku 4.5',
+    family: 'claude-haiku',
+    attachment: true,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 144000, output: 32768 },
+  },
+  'gemini-3.1-pro-preview': {
+    id: 'gemini-3.1-pro-preview',
+    name: 'Gemini 3.1 Pro Preview',
+    family: 'gemini-pro',
+    attachment: true,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image', 'audio'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 128000, output: 32768 },
+  },
+  'gemini-3-flash-preview': {
+    id: 'gemini-3-flash-preview',
+    name: 'Gemini 3 Flash',
+    family: 'gemini-flash',
+    attachment: true,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 128000, output: 32768 },
+  },
+  'gemini-2.5-pro': {
+    id: 'gemini-2.5-pro',
+    name: 'Gemini 2.5 Pro',
+    family: 'gemini-pro',
+    attachment: true,
+    reasoning: false,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text', 'image'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 128000, output: 32768 },
+  },
+  'grok-code-fast-1': {
+    id: 'grok-code-fast-1',
+    name: 'Grok Code Fast 1',
+    family: 'grok',
+    attachment: false,
+    reasoning: true,
+    tool_call: true,
+    temperature: true,
+    knowledge: '2025-05',
+    release_date: '2025-05-01',
+    last_updated: '2025-05-01',
+    modalities: { input: ['text'], output: ['text'] },
+    open_weights: false,
+    cost: { input: 0, output: 0 },
+    limit: { context: 128000, output: 32768 },
+  },
+}
+
+export function getCopilotModelIds(): string[] {
+  return Object.keys(COPILOT_MODELS)
+}
+
+export function getCopilotModel(id: string): CopilotModel | undefined {
+  return COPILOT_MODELS[id]
+}
+
+export function getAllCopilotModels(): CopilotModel[] {
+  return Object.values(COPILOT_MODELS)
+}

src/utils/model/model.ts

@@ -43,6 +43,10 @@ export function getSmallFastModel(): ModelName {
   if (getAPIProvider() === 'openai') {
     return process.env.OPENAI_MODEL || 'gpt-4o-mini'
   }
+  // For GitHub Copilot provider
+  if (getAPIProvider() === 'github') {
+    return process.env.OPENAI_MODEL || 'github:copilot'
+  }
   return getDefaultHaikuModel()
 }
 
@@ -137,6 +141,10 @@ export function getDefaultOpusModel(): ModelName {
   if (getAPIProvider() === 'codex') {
     return process.env.OPENAI_MODEL || 'gpt-5.4'
   }
+  // GitHub Copilot provider
+  if (getAPIProvider() === 'github') {
+    return process.env.OPENAI_MODEL || 'github:copilot'
+  }
   // 3P providers (Bedrock, Vertex, Foundry) — kept as a separate branch
   // even when values match, since 3P availability lags firstParty and
   // these will diverge again at the next model launch.
@@ -163,6 +171,10 @@ export function getDefaultSonnetModel(): ModelName {
   if (getAPIProvider() === 'codex') {
     return process.env.OPENAI_MODEL || 'gpt-5.4'
   }
+  // GitHub Copilot provider
+  if (getAPIProvider() === 'github') {
+    return process.env.OPENAI_MODEL || 'github:copilot'
+  }
   // Default to Sonnet 4.5 for 3P since they may not have 4.6 yet
   if (getAPIProvider() !== 'firstParty') {
     return getModelStrings().sonnet45
@@ -175,10 +187,6 @@ export function getDefaultHaikuModel(): ModelName {
   if (process.env.ANTHROPIC_DEFAULT_HAIKU_MODEL) {
     return process.env.ANTHROPIC_DEFAULT_HAIKU_MODEL
   }
-  // Gemini provider
-  if (getAPIProvider() === 'gemini') {
-    return process.env.GEMINI_MODEL || 'gemini-2.0-flash-lite'
-  }
   // OpenAI provider
   if (getAPIProvider() === 'openai') {
     return process.env.OPENAI_MODEL || 'gpt-4o-mini'
@@ -187,6 +195,14 @@ export function getDefaultHaikuModel(): ModelName {
   if (getAPIProvider() === 'codex') {
     return process.env.OPENAI_MODEL || 'gpt-5.4'
   }
+  // GitHub Copilot provider
+  if (getAPIProvider() === 'github') {
+    return process.env.OPENAI_MODEL || 'github:copilot'
+  }
+  // Gemini provider
+  if (getAPIProvider() === 'gemini') {
+    return process.env.GEMINI_MODEL || 'gemini-2.0-flash-lite'
+  }
 
   // Haiku 4.5 is available on all platforms (first-party, Foundry, Bedrock, Vertex)
   return getModelStrings().haiku45
@@ -231,6 +247,11 @@ export function getRuntimeMainLoopModel(params: {
  * @returns The default model setting to use
  */
 export function getDefaultMainLoopModelSetting(): ModelName | ModelAlias {
+  // GitHub Copilot provider: check settings.model first, then env, then default
+  if (getAPIProvider() === 'github') {
+    const settings = getSettings_DEPRECATED() || {}
+    return settings.model || process.env.OPENAI_MODEL || 'github:copilot'
+  }
   // Gemini provider: always use the configured Gemini model
   if (getAPIProvider() === 'gemini') {
     return process.env.GEMINI_MODEL || 'gemini-2.0-flash'
@@ -239,10 +260,6 @@ export function getDefaultMainLoopModelSetting(): ModelName | ModelAlias {
   if (getAPIProvider() === 'openai') {
     return process.env.OPENAI_MODEL || 'gpt-4o'
   }
-  // GitHub provider: always use the configured GitHub model
-  if (getAPIProvider() === 'github') {
-    return process.env.OPENAI_MODEL || 'github:copilot'
-  }
   // Codex provider: always use the configured Codex model (default gpt-5.4)
   if (getAPIProvider() === 'codex') {
     return process.env.OPENAI_MODEL || 'gpt-5.4'
@@ -426,8 +443,33 @@ export function renderModelSetting(setting: ModelName | ModelAlias): string {
  * if the model is not recognized as a public model.
  */
 export function getPublicModelDisplayName(model: ModelName): string | null {
-  // For OpenAI/Gemini/Codex providers, show the actual model name not a Claude alias
+  // For OpenAI/Gemini/Codex/GitHub providers, show the actual model name not a Claude alias
-  if (getAPIProvider() === 'openai' || getAPIProvider() === 'gemini' || getAPIProvider() === 'codex') {
+  if (getAPIProvider() === 'openai' || getAPIProvider() === 'gemini' || getAPIProvider() === 'codex' || getAPIProvider() === 'github') {
+    // Return display names for known GitHub Copilot models
+    const copilotModelNames: Record<string, string> = {
+      'gpt-5.4': 'GPT-5.4',
+      'gpt-5.4-mini': 'GPT-5.4 mini',
+      'gpt-5.3-codex': 'GPT-5.3 Codex',
+      'gpt-5.2-codex': 'GPT-5.2 Codex',
+      'gpt-5.2': 'GPT-5.2',
+      'gpt-5.1-codex': 'GPT-5.1 Codex',
+      'gpt-5.1-codex-max': 'GPT-5.1 Codex max',
+      'gpt-5.1-codex-mini': 'GPT-5.1 Codex mini',
+      'gpt-4o': 'GPT-4o',
+      'gpt-4.1': 'GPT-4.1',
+      'claude-opus-4.6': 'Claude Opus 4.6',
+      'claude-opus-4.5': 'Claude Opus 4.5',
+      'claude-sonnet-4.6': 'Claude Sonnet 4.6',
+      'claude-sonnet-4.5': 'Claude Sonnet 4.5',
+      'claude-haiku-4.5': 'Claude Haiku 4.5',
+      'gemini-3.1-pro-preview': 'Gemini 3.1 Pro Preview',
+      'gemini-3-flash-preview': 'Gemini 3 Flash',
+      'gemini-2.5-pro': 'Gemini 2.5 Pro',
+      'grok-code-fast-1': 'Grok Code Fast 1',
+    }
+    if (copilotModelNames[model]) {
+      return copilotModelNames[model]
+    }
     return null
   }
   switch (model) {
@@ -484,6 +526,10 @@ export function renderModelName(model: ModelName): string {
   if (publicName) {
     return publicName
   }
+  // Handle GitHub Copilot special model aliases
+  if (model === 'github:copilot') {
+    return 'GPT-4o'
+  }
   if (process.env.USER_TYPE === 'ant') {
     const resolved = parseUserSpecifiedModel(model)
     const antModel = resolveAntModel(model)

src/utils/model/modelOptions.github.test.ts

@@ -1,7 +1,17 @@
-import { afterEach, expect, mock, test } from 'bun:test'
+import { afterEach, beforeEach, expect, mock, test } from 'bun:test'
 
+import { resetModelStringsForTestingOnly } from '../../bootstrap/state.js'
 import { saveGlobalConfig } from '../config.js'
 
+async function importFreshModelOptionsModule() {
+  mock.restore()
+  mock.module('./providers.js', () => ({
+    getAPIProvider: () => 'github',
+  }))
+  const nonce = `${Date.now()}-${Math.random()}`
+  return import(`./modelOptions.js?ts=${nonce}`)
+}
+
 const originalEnv = {
   CLAUDE_CODE_USE_GITHUB: process.env.CLAUDE_CODE_USE_GITHUB,
   CLAUDE_CODE_USE_OPENAI: process.env.CLAUDE_CODE_USE_OPENAI,
@@ -14,6 +24,20 @@ const originalEnv = {
   ANTHROPIC_CUSTOM_MODEL_OPTION: process.env.ANTHROPIC_CUSTOM_MODEL_OPTION,
 }
 
+beforeEach(() => {
+  mock.restore()
+  delete process.env.CLAUDE_CODE_USE_GITHUB
+  delete process.env.CLAUDE_CODE_USE_OPENAI
+  delete process.env.CLAUDE_CODE_USE_GEMINI
+  delete process.env.CLAUDE_CODE_USE_BEDROCK
+  delete process.env.CLAUDE_CODE_USE_VERTEX
+  delete process.env.CLAUDE_CODE_USE_FOUNDRY
+  delete process.env.OPENAI_MODEL
+  delete process.env.OPENAI_BASE_URL
+  delete process.env.ANTHROPIC_CUSTOM_MODEL_OPTION
+  resetModelStringsForTestingOnly()
+})
+
 afterEach(() => {
   process.env.CLAUDE_CODE_USE_GITHUB = originalEnv.CLAUDE_CODE_USE_GITHUB
   process.env.CLAUDE_CODE_USE_OPENAI = originalEnv.CLAUDE_CODE_USE_OPENAI
@@ -34,18 +58,10 @@ afterEach(() => {
     providerProfiles: [],
     activeProviderProfileId: undefined,
   }))
+  resetModelStringsForTestingOnly()
 })
 
-async function importFreshModelOptionsModule() {
+test('GitHub provider exposes default + all Copilot models in /model options', async () => {
-  mock.restore()
-  mock.module('./providers.js', () => ({
-    getAPIProvider: () => 'github',
-  }))
-  const nonce = `${Date.now()}-${Math.random()}`
-  return import(`./modelOptions.js?ts=${nonce}`)
-}
-
-test('GitHub provider exposes only default + GitHub model in /model options', async () => {
   process.env.CLAUDE_CODE_USE_GITHUB = '1'
   delete process.env.CLAUDE_CODE_USE_OPENAI
   delete process.env.CLAUDE_CODE_USE_GEMINI
@@ -53,13 +69,16 @@ test('GitHub provider exposes only default + GitHub model in /model options', as
   delete process.env.CLAUDE_CODE_USE_VERTEX
   delete process.env.CLAUDE_CODE_USE_FOUNDRY
 
-  process.env.OPENAI_MODEL = 'github:copilot'
+  process.env.OPENAI_MODEL = 'gpt-4o'
   delete process.env.ANTHROPIC_CUSTOM_MODEL_OPTION
 
   const { getModelOptions } = await importFreshModelOptionsModule()
   const options = getModelOptions(false)
-  const nonDefault = options.filter(option => option.value !== null)
+  const nonDefault = options.filter(
+    (option: { value: unknown }) => option.value !== null,
+  )
 
-  expect(nonDefault.length).toBe(1)
+  expect(nonDefault.length).toBeGreaterThan(1)
-  expect(nonDefault[0]?.value).toBe('github:copilot')
+  expect(nonDefault.some((o: { value: unknown }) => o.value === 'gpt-4o')).toBe(true)
+  expect(nonDefault.some((o: { value: unknown }) => o.value === 'gpt-5.3-codex')).toBe(true)
 })

src/utils/model/modelOptions.ts

@@ -35,6 +35,7 @@ import { has1mContext } from '../context.js'
 import { getGlobalConfig } from '../config.js'
 import { getActiveOpenAIModelOptionsCache } from '../providerProfiles.js'
 import { getCachedOllamaModelOptions, isOllamaProvider } from './ollamaModels.js'
+import { getAntModels } from './antModels.js'
 
 // @[MODEL LAUNCH]: Update all the available and default model option strings below.
 
@@ -351,17 +352,20 @@ function getCodexModelOptions(): ModelOption[] {
 
 // @[MODEL LAUNCH]: Update the model picker lists below to include/reorder options for the new model.
 // Each user tier (ant, Max/Team Premium, Pro/Team Standard/Enterprise, PAYG 1P, PAYG 3P) has its own list.
+
+import { getAllCopilotModels } from './copilotModels.js'
+
+function getCopilotModelOptions(): ModelOption[] {
+  return getAllCopilotModels().map(m => ({
+    value: m.id,
+    label: m.name,
+    description: `${m.family}${m.reasoning ? ' · Reasoning' : ''}${m.tool_call ? ' · Tool call' : ''} · ${Math.round(m.limit.context / 1000)}K context`,
+  }))
+}
+
 function getModelOptionsBase(fastMode = false): ModelOption[] {
   if (getAPIProvider() === 'github') {
-    const githubModel = process.env.OPENAI_MODEL?.trim() || 'github:copilot'
+    return [getDefaultOptionForUser(fastMode), ...getCopilotModelOptions()]
-    return [
-      getDefaultOptionForUser(fastMode),
-      {
-        value: githubModel,
-        label: githubModel,
-        description: 'GitHub Models default',
-      },
-    ]
   }
 
   // When using Ollama, show models from the Ollama server instead of Claude models

src/utils/model/openaiContextWindows.ts

@@ -44,6 +44,10 @@ const OPENAI_CONTEXT_WINDOWS: Record<string, number> = {
   'mistral-large-latest':     131_072,
   'mistral-small-latest':     131_072,
 
+  // MiniMax
+  'MiniMax-M2.7':             204_800,
+  'minimax-m2.7':             204_800,
+
   // Google (via OpenRouter)
   'google/gemini-2.0-flash':1_048_576,
   'google/gemini-2.5-pro':  1_048_576,
@@ -110,6 +114,10 @@ const OPENAI_MAX_OUTPUT_TOKENS: Record<string, number> = {
   'mistral-large-latest':     32_768,
   'mistral-small-latest':     32_768,
 
+  // MiniMax
+  'MiniMax-M2.7':            131_072,
+  'minimax-m2.7':            131_072,
+
   // Google (via OpenRouter)
   'google/gemini-2.0-flash':   8_192,
   'google/gemini-2.5-pro':    65_536,

src/utils/permissions/dangerousPatterns.ts

@@ -51,6 +51,7 @@ export const DANGEROUS_BASH_PATTERNS: readonly string[] = [
   'xargs',
   'sudo',
   // Internal-only: internal-only tools plus general tools that ant sandbox
+  // data shows are frequently over-allowlisted as broad prefixes.
   // dotfile data shows are commonly over-allowlisted as broad prefixes.
   // These stay internal-only — external users don't have coo, and the rest are
   // an empirical-risk call grounded in ant sandbox data, not a universal

src/utils/providerFlag.test.ts

@@ -1,4 +1,4 @@
-import { describe, expect, test, afterEach } from 'bun:test'
+import { afterEach, beforeEach, describe, expect, test } from 'bun:test'
 import {
   parseProviderFlag,
   applyProviderFlag,
@@ -6,10 +6,7 @@ import {
   VALID_PROVIDERS,
 } from './providerFlag.js'
 
-const originalEnv = { ...process.env }
+const ENV_KEYS = [
-
-afterEach(() => {
-  for (const key of [
   'CLAUDE_CODE_USE_OPENAI',
   'CLAUDE_CODE_USE_GEMINI',
   'CLAUDE_CODE_USE_GITHUB',
@@ -19,9 +16,42 @@ afterEach(() => {
   'OPENAI_API_KEY',
   'OPENAI_MODEL',
   'GEMINI_MODEL',
-  ]) {
+]
-    if (originalEnv[key] === undefined) delete process.env[key]
+
-    else process.env[key] = originalEnv[key]
+const originalEnv: Record<string, string | undefined> = {}
+
+beforeEach(() => {
+  for (const key of ENV_KEYS) {
+    originalEnv[key] = process.env[key]
+    delete process.env[key]
+  }
+})
+
+const RESET_KEYS = [
+  'CLAUDE_CODE_USE_OPENAI',
+  'CLAUDE_CODE_USE_GEMINI',
+  'CLAUDE_CODE_USE_GITHUB',
+  'CLAUDE_CODE_USE_BEDROCK',
+  'CLAUDE_CODE_USE_VERTEX',
+  'OPENAI_BASE_URL',
+  'OPENAI_API_KEY',
+  'OPENAI_MODEL',
+  'GEMINI_MODEL',
+] as const
+
+beforeEach(() => {
+  for (const key of RESET_KEYS) {
+    delete process.env[key]
+  }
+})
+
+afterEach(() => {
+  for (const key of ENV_KEYS) {
+    if (originalEnv[key] === undefined) {
+      delete process.env[key]
+    } else {
+      process.env[key] = originalEnv[key]
+    }
   }
 })
 

src/utils/providerProfiles.test.ts

@@ -2,6 +2,10 @@ import { afterEach, describe, expect, mock, test } from 'bun:test'
 
 import type { ProviderProfile } from './config.js'
 
+async function importFreshProvidersModule() {
+  return import(`./model/providers.ts?ts=${Date.now()}-${Math.random()}`)
+}
+
 const originalEnv = { ...process.env }
 
 const RESTORED_KEYS = [
@@ -96,24 +100,26 @@ function buildProfile(overrides: Partial<ProviderProfile> = {}): ProviderProfile
 
 describe('applyProviderProfileToProcessEnv', () => {
   test('openai profile clears competing gemini/github flags', async () => {
-    const { applyProviderProfileToProcessEnv, getAPIProvider } =
+    const { applyProviderProfileToProcessEnv } =
       await importFreshProviderProfileModules()
     process.env.CLAUDE_CODE_USE_GEMINI = '1'
     process.env.CLAUDE_CODE_USE_GITHUB = '1'
 
     applyProviderProfileToProcessEnv(buildProfile())
+    const { getAPIProvider: getFreshAPIProvider } =
+      await importFreshProvidersModule()
 
     expect(process.env.CLAUDE_CODE_USE_GEMINI).toBeUndefined()
     expect(process.env.CLAUDE_CODE_USE_GITHUB).toBeUndefined()
-    expect(process.env.CLAUDE_CODE_USE_OPENAI).toBe('1')
+    expect(String(process.env.CLAUDE_CODE_USE_OPENAI)).toBe('1')
     expect(process.env.CLAUDE_CODE_PROVIDER_PROFILE_ENV_APPLIED_ID).toBe(
       'provider_test',
     )
-    expect(getAPIProvider()).toBe('openai')
+    expect(getFreshAPIProvider()).toBe('openai')
   })
 
   test('anthropic profile clears competing gemini/github flags', async () => {
-    const { applyProviderProfileToProcessEnv, getAPIProvider } =
+    const { applyProviderProfileToProcessEnv } =
       await importFreshProviderProfileModules()
     process.env.CLAUDE_CODE_USE_GEMINI = '1'
     process.env.CLAUDE_CODE_USE_GITHUB = '1'
@@ -125,11 +131,13 @@ describe('applyProviderProfileToProcessEnv', () => {
         model: 'claude-sonnet-4-6',
       }),
     )
+    const { getAPIProvider: getFreshAPIProvider } =
+      await importFreshProvidersModule()
 
     expect(process.env.CLAUDE_CODE_USE_GEMINI).toBeUndefined()
     expect(process.env.CLAUDE_CODE_USE_GITHUB).toBeUndefined()
     expect(process.env.CLAUDE_CODE_USE_OPENAI).toBeUndefined()
-    expect(getAPIProvider()).toBe('firstParty')
+    expect(getFreshAPIProvider()).toBe('firstParty')
   })
 })
 
@@ -177,7 +185,7 @@ describe('applyActiveProviderProfileFromConfig', () => {
     } as any)
 
     expect(applied).toBeUndefined()
-    expect(process.env.CLAUDE_CODE_USE_OPENAI).toBe('1')
+    expect(String(process.env.CLAUDE_CODE_USE_OPENAI)).toBe('1')
     expect(process.env.OPENAI_BASE_URL).toBe('http://localhost:11434/v1')
     expect(process.env.OPENAI_MODEL).toBe('qwen2.5:3b')
   })
@@ -267,7 +275,7 @@ describe('applyActiveProviderProfileFromConfig', () => {
     } as any)
 
     expect(applied?.id).toBe('saved_openai')
-    expect(process.env.CLAUDE_CODE_USE_OPENAI).toBe('1')
+    expect(String(process.env.CLAUDE_CODE_USE_OPENAI)).toBe('1')
     expect(process.env.OPENAI_BASE_URL).toBe('https://api.openai.com/v1')
     expect(process.env.OPENAI_MODEL).toBe('gpt-4o')
   })
@@ -303,7 +311,7 @@ describe('persistActiveProviderProfileModel', () => {
     )
 
     const saved = getProviderProfiles().find(
-      profile => profile.id === activeProfile.id,
+      (profile: ProviderProfile) => profile.id === activeProfile.id,
     )
     expect(saved?.model).toBe('minimax-m2.5:cloud')
   })
@@ -333,7 +341,7 @@ describe('persistActiveProviderProfileModel', () => {
 
     expect(process.env.OPENAI_MODEL).toBe('cli-model')
     const saved = getProviderProfiles().find(
-      profile => profile.id === activeProfile.id,
+      (profile: ProviderProfile) => profile.id === activeProfile.id,
     )
     expect(saved?.model).toBe('minimax-m2.5:cloud')
   })
@@ -414,7 +422,7 @@ describe('deleteProviderProfile', () => {
     expect(result.activeProfileId).toBeUndefined()
 
     expect(process.env.CLAUDE_CODE_PROVIDER_PROFILE_ENV_APPLIED).toBeUndefined()
-    expect(process.env.CLAUDE_CODE_USE_OPENAI).toBe('1')
+    expect(String(process.env.CLAUDE_CODE_USE_OPENAI)).toBe('1')
     expect(process.env.OPENAI_BASE_URL).toBe('http://localhost:11434/v1')
     expect(process.env.OPENAI_MODEL).toBe('qwen2.5:3b')
   })

src/utils/providerValidation.ts

@@ -1,4 +1,5 @@
 import {
+  getGithubEndpointType,
   isLocalProviderUrl,
   resolveCodexApiCredentials,
   resolveProviderRequest,
@@ -15,6 +16,51 @@ function isEnvTruthy(value: string | undefined): boolean {
   return normalized !== '' && normalized !== '0' && normalized !== 'false' && normalized !== 'no'
 }
 
+type GithubTokenStatus = 'valid' | 'expired' | 'invalid_format'
+
+const GITHUB_PAT_PREFIXES = ['ghp_', 'gho_', 'ghs_', 'ghr_', 'github_pat_']
+
+function checkGithubTokenStatus(
+  token: string,
+  endpointType: 'copilot' | 'models' | 'custom' = 'copilot',
+): GithubTokenStatus {
+  // PATs work with GitHub Models but not with Copilot API
+  if (GITHUB_PAT_PREFIXES.some(prefix => token.startsWith(prefix))) {
+    if (endpointType === 'copilot') {
+      return 'expired'
+    }
+    return 'valid'
+  }
+
+  const expMatch = token.match(/exp=(\d+)/)
+  if (expMatch) {
+    const expSeconds = Number(expMatch[1])
+    if (!Number.isNaN(expSeconds)) {
+      return Date.now() >= expSeconds * 1000 ? 'expired' : 'valid'
+    }
+  }
+
+  const parts = token.split('.')
+  const looksLikeJwt =
+    parts.length === 3 && parts.every(part => /^[A-Za-z0-9_-]+$/.test(part))
+  if (looksLikeJwt) {
+    try {
+      const normalized = parts[1].replace(/-/g, '+').replace(/_/g, '/')
+      const padded = normalized + '='.repeat((4 - (normalized.length % 4)) % 4)
+      const json = Buffer.from(padded, 'base64').toString('utf8')
+      const parsed = JSON.parse(json)
+      if (parsed && typeof parsed === 'object' && parsed.exp) {
+        return Date.now() >= (parsed.exp as number) * 1000 ? 'expired' : 'valid'
+      }
+    } catch {
+      return 'invalid_format'
+    }
+  }
+
+  // Keep compatibility with opaque token formats that do not expose expiry.
+  return 'valid'
+}
+
 export async function getProviderValidationError(
   env: NodeJS.ProcessEnv = process.env,
   options?: {
@@ -39,7 +85,19 @@ export async function getProviderValidationError(
   if (useGithub && !useOpenAI) {
     const token = (env.GITHUB_TOKEN?.trim() || env.GH_TOKEN?.trim()) ?? ''
     if (!token) {
-      return 'GITHUB_TOKEN or GH_TOKEN is required when CLAUDE_CODE_USE_GITHUB=1.'
+      return 'GitHub Copilot authentication required.\n' +
+        'Run /onboard-github in the CLI to sign in with your GitHub account.\n' +
+        'This will store your OAuth token securely and enable Copilot models.'
+    }
+    const endpointType = getGithubEndpointType(env.OPENAI_BASE_URL)
+    const status = checkGithubTokenStatus(token, endpointType)
+    if (status === 'expired') {
+      return 'GitHub Copilot token has expired.\n' +
+        'Run /onboard-github to sign in again and get a fresh token.'
+    }
+    if (status === 'invalid_format') {
+      return 'GitHub Copilot token is invalid or corrupted.\n' +
+        'Run /onboard-github to sign in again with your GitHub account.'
     }
     return null
   }

src/utils/ripgrep.test.ts

@@ -1,11 +1,52 @@
 import { expect, test } from 'bun:test'
+import path from 'path'
 
-import { wrapRipgrepUnavailableError } from './ripgrep.ts'
+import { resolveRipgrepConfig, wrapRipgrepUnavailableError } from './ripgrep.js'
+
+const MOCK_BUILTIN_PATH = path.normalize(
+  process.platform === 'win32'
+    ? `vendor/ripgrep/${process.arch}-win32/rg.exe`
+    : `vendor/ripgrep/${process.arch}-${process.platform}/rg`,
+)
+
+test('ripgrepCommand falls back to system rg when builtin binary is missing', () => {
+  const config = resolveRipgrepConfig({
+    userWantsSystemRipgrep: false,
+    bundledMode: false,
+    builtinCommand: MOCK_BUILTIN_PATH,
+    builtinExists: false,
+    systemExecutablePath: '/usr/bin/rg',
+    processExecPath: '/fake/bun',
+  })
+
+  expect(config).toMatchObject({
+    mode: 'system',
+    command: 'rg',
+    args: [],
+  })
+})
+
+test('ripgrepCommand keeps builtin mode when bundled binary exists', () => {
+  const config = resolveRipgrepConfig({
+    userWantsSystemRipgrep: false,
+    bundledMode: false,
+    builtinCommand: MOCK_BUILTIN_PATH,
+    builtinExists: true,
+    systemExecutablePath: '/usr/bin/rg',
+    processExecPath: '/fake/bun',
+  })
+
+  expect(config).toMatchObject({
+    mode: 'builtin',
+    command: MOCK_BUILTIN_PATH,
+    args: [],
+  })
+})
 
 test('wrapRipgrepUnavailableError explains missing packaged fallback', () => {
   const error = wrapRipgrepUnavailableError(
     { code: 'ENOENT', message: 'spawn rg ENOENT' },
-    { mode: 'builtin', command: 'C:\\fake\\vendor\\ripgrep\\rg.exe' },
+    { mode: 'builtin', command: 'C:\\fake\\vendor\\ripgrep\\rg.exe', args: [] },
     'win32',
   )
 
@@ -18,7 +59,7 @@ test('wrapRipgrepUnavailableError explains missing packaged fallback', () => {
 test('wrapRipgrepUnavailableError explains missing system ripgrep', () => {
   const error = wrapRipgrepUnavailableError(
     { code: 'ENOENT', message: 'spawn rg ENOENT' },
-    { mode: 'system', command: 'rg' },
+    { mode: 'system', command: 'rg', args: [] },
     'linux',
   )
 

src/utils/ripgrep.ts

@@ -1,5 +1,6 @@
 import type { ChildProcess, ExecFileException } from 'child_process'
 import { execFile, spawn } from 'child_process'
+import { existsSync } from 'fs'
 import memoize from 'lodash-es/memoize.js'
 import { homedir } from 'os'
 import * as path from 'path'
@@ -30,40 +31,72 @@ type RipgrepConfig = {
 
 type RipgrepErrorLike = Pick<NodeJS.ErrnoException, 'code' | 'message'>
 
-const getRipgrepConfig = memoize((): RipgrepConfig => {
+function isErrnoException(error: unknown): error is NodeJS.ErrnoException {
-  const userWantsSystemRipgrep = isEnvDefinedFalsy(
+  return error instanceof Error
-    process.env.USE_BUILTIN_RIPGREP,
+}
-  )
 
-  // Try system ripgrep if user wants it
+type ResolveRipgrepConfigArgs = {
-  if (userWantsSystemRipgrep) {
+  userWantsSystemRipgrep: boolean
-    const { cmd: systemPath } = findExecutable('rg', [])
+  bundledMode: boolean
-    if (systemPath !== 'rg') {
+  builtinCommand: string
-      // SECURITY: Use command name 'rg' instead of systemPath to prevent PATH hijacking
+  builtinExists: boolean
-      // If we used systemPath, a malicious ./rg.exe in current directory could be executed
+  systemExecutablePath: string
-      // Using just 'rg' lets the OS resolve it safely with NoDefaultCurrentDirectoryInExePath protection
+  processExecPath?: string
+}
+
+export function resolveRipgrepConfig({
+  userWantsSystemRipgrep,
+  bundledMode,
+  builtinCommand,
+  builtinExists,
+  systemExecutablePath,
+  processExecPath = process.execPath,
+}: ResolveRipgrepConfigArgs): RipgrepConfig {
+  if (userWantsSystemRipgrep && systemExecutablePath !== 'rg') {
+    // SECURITY: Use command name 'rg' instead of systemExecutablePath to prevent PATH hijacking
     return { mode: 'system', command: 'rg', args: [] }
   }
-  }
 
-  // In bundled (native) mode, ripgrep is statically compiled into bun-internal
+  if (bundledMode) {
-  // and dispatches based on argv[0]. We spawn ourselves with argv0='rg'.
-  if (isInBundledMode()) {
     return {
       mode: 'embedded',
-      command: process.execPath,
+      command: processExecPath,
       args: ['--no-config'],
       argv0: 'rg',
     }
   }
 
+  if (builtinExists) {
+    return { mode: 'builtin', command: builtinCommand, args: [] }
+  }
+
+  if (systemExecutablePath !== 'rg') {
+    return { mode: 'system', command: 'rg', args: [] }
+  }
+
+  return { mode: 'builtin', command: builtinCommand, args: [] }
+}
+
+const getRipgrepConfig = memoize((): RipgrepConfig => {
+  const userWantsSystemRipgrep = isEnvDefinedFalsy(
+    process.env.USE_BUILTIN_RIPGREP,
+  )
+  const bundledMode = isInBundledMode()
   const rgRoot = path.resolve(__dirname, 'vendor', 'ripgrep')
-  const command =
+  const builtinCommand =
     process.platform === 'win32'
       ? path.resolve(rgRoot, `${process.arch}-win32`, 'rg.exe')
       : path.resolve(rgRoot, `${process.arch}-${process.platform}`, 'rg')
+  const builtinExists = existsSync(builtinCommand)
+  const { cmd: systemExecutablePath } = findExecutable('rg', [])
 
-  return { mode: 'builtin', command, args: [] }
+  return resolveRipgrepConfig({
+    userWantsSystemRipgrep,
+    bundledMode,
+    builtinCommand,
+    builtinExists,
+    systemExecutablePath,
+  })
 })
 
 export function ripgrepCommand(): {
@@ -324,7 +357,9 @@ async function ripGrepFileCount(
       if (settled) return
       settled = true
       reject(
-        err.code === 'ENOENT' ? wrapRipgrepUnavailableError(err) : err,
+        isErrnoException(err) && err.code === 'ENOENT'
+          ? wrapRipgrepUnavailableError(err)
+          : err,
       )
     })
   })
@@ -388,7 +423,9 @@ export async function ripGrepStream(
       if (settled) return
       settled = true
       reject(
-        err.code === 'ENOENT' ? wrapRipgrepUnavailableError(err) : err,
+        isErrnoException(err) && err.code === 'ENOENT'
+          ? wrapRipgrepUnavailableError(err)
+          : err,
       )
     })
   })
@@ -436,7 +473,9 @@ export async function ripGrep(
       const CRITICAL_ERROR_CODES = ['ENOENT', 'EACCES', 'EPERM']
       if (CRITICAL_ERROR_CODES.includes(error.code as string)) {
         reject(
-          error.code === 'ENOENT' ? wrapRipgrepUnavailableError(error) : error,
+          isErrnoException(error) && error.code === 'ENOENT'
+            ? wrapRipgrepUnavailableError(error)
+            : error,
         )
         return
       }

src/utils/schemaSanitizer.test.ts

@@ -0,0 +1,68 @@
+import { describe, expect, test } from 'bun:test'
+
+import { sanitizeSchemaForOpenAICompat } from './schemaSanitizer'
+
+describe('sanitizeSchemaForOpenAICompat', () => {
+  test('preserves Grep-like properties.pattern while keeping it required', () => {
+    const schema = {
+      type: 'object',
+      properties: {
+        pattern: {
+          type: 'string',
+          description: 'The regular expression pattern to search for in file contents',
+        },
+        path: { type: 'string' },
+        glob: { type: 'string' },
+      },
+      required: ['pattern'],
+    }
+
+    const sanitized = sanitizeSchemaForOpenAICompat(schema)
+    const properties = sanitized.properties as Record<string, unknown> | undefined
+
+    expect(Object.keys(properties ?? {})).toEqual(['pattern', 'path', 'glob'])
+    expect(properties?.pattern).toEqual({
+      type: 'string',
+      description: 'The regular expression pattern to search for in file contents',
+    })
+    expect(sanitized.required).toEqual(['pattern'])
+  })
+
+  test('preserves Glob-like properties.pattern while keeping it required', () => {
+    const schema = {
+      type: 'object',
+      properties: {
+        pattern: {
+          type: 'string',
+          description: 'The glob pattern to match files against',
+        },
+        path: { type: 'string' },
+      },
+      required: ['pattern'],
+    }
+
+    const sanitized = sanitizeSchemaForOpenAICompat(schema)
+    const properties = sanitized.properties as Record<string, unknown> | undefined
+
+    expect(Object.keys(properties ?? {})).toEqual(['pattern', 'path'])
+    expect(properties?.pattern).toEqual({
+      type: 'string',
+      description: 'The glob pattern to match files against',
+    })
+    expect(sanitized.required).toEqual(['pattern'])
+  })
+
+  test('strips JSON Schema validator pattern from string schemas', () => {
+    const schema = {
+      type: 'string',
+      pattern: '^[a-z]+$',
+      minLength: 1,
+    }
+
+    const sanitized = sanitizeSchemaForOpenAICompat(schema)
+
+    expect(sanitized).toEqual({
+      type: 'string',
+    })
+  })
+})

src/utils/schemaSanitizer.ts

@@ -33,6 +33,15 @@ function stripSchemaKeywords(schema: unknown, keywords: Set<string>): unknown {
 
   const result: Record<string, unknown> = {}
   for (const [key, value] of Object.entries(schema)) {
+    if (key === 'properties' && isSchemaRecord(value)) {
+      const sanitizedProps: Record<string, unknown> = {}
+      for (const [propName, propSchema] of Object.entries(value)) {
+        sanitizedProps[propName] = stripSchemaKeywords(propSchema, keywords)
+      }
+      result[key] = sanitizedProps
+      continue
+    }
+
     if (keywords.has(key)) {
       continue
     }
@@ -215,10 +224,13 @@ export function sanitizeSchemaForOpenAICompat(
     }
   }
 
-  if (Array.isArray(record.required) && isSchemaRecord(record.properties)) {
+  const properties = isSchemaRecord(record.properties)
+    ? record.properties
+    : undefined
+
+  if (Array.isArray(record.required) && properties) {
     record.required = record.required.filter(
-      (value): value is string =>
+      (value): value is string => typeof value === 'string' && value in properties,
-        typeof value === 'string' && value in record.properties,
     )
   }
 

src/utils/settings/types.ts

@@ -27,6 +27,7 @@ export {
 
 // Also import for use within this file
 import { type HookCommand, HooksSchema } from '../../schemas/hooks.js'
+import { AutoFixConfigSchema } from '../../services/autoFix/autoFixConfig.js'
 import { count } from '../array.js'
 
 /**
@@ -435,6 +436,12 @@ export const SettingsSchema = lazySchema(() =>
       hooks: HooksSchema()
         .optional()
         .describe('Custom commands to run before/after tool executions'),
+      autoFix: AutoFixConfigSchema
+        .optional()
+        .describe(
+          'Auto-fix configuration: automatically run lint/test after AI file edits ' +
+          'and feed errors back for self-repair.',
+        ),
       worktree: z
         .object({
           symlinkDirectories: z

src/utils/staticRender.tsx

@@ -97,8 +97,12 @@ export function renderToAnsiString(node: React.ReactNode, columns?: number): Pro
       patchConsole: false
     });
 
-    // Wait for the component to exit naturally
+    // Wait for the component to exit naturally, with a timeout guard so
-    await instance.waitUntilExit();
+    // tests never hang indefinitely if a render error prevents exit().
+    await Promise.race([
+      instance.waitUntilExit(),
+      new Promise<void>(resolve => setTimeout(resolve, 3000)),
+    ]);
 
     // Extract only the first frame's content to avoid duplication
     // (Ink outputs multiple frames in non-TTY mode)