feat(codex): add dashboard control center

- add guided editors for top-level settings, trust, profiles, providers, MCP, and features

- refresh raw snapshots after patch saves to avoid stale mtime conflicts

- block structured saves while raw TOML is dirty and add route plus hook coverage

ui/tests/unit/hooks/use-codex.test.tsx

@@ -0,0 +1,142 @@
+import { act, renderHook, waitFor } from '@testing-library/react';
+import { afterEach, describe, expect, it, vi } from 'vitest';
+import type { ReactNode } from 'react';
+import { AllProviders } from '../../setup/test-utils';
+import { useCodex } from '@/hooks/use-codex';
+
+function createJsonResponse(body: Record<string, unknown>, status = 200): Response {
+  return new Response(JSON.stringify(body), {
+    status,
+    headers: { 'Content-Type': 'application/json' },
+  });
+}
+
+const diagnosticsResponse = {
+  binary: {
+    installed: true,
+    path: '/tmp/codex',
+    installDir: '/tmp',
+    source: 'PATH',
+    version: 'codex-cli 0.118.0-alpha.3',
+    overridePath: null,
+    supportsConfigOverrides: true,
+  },
+  file: {
+    label: 'Codex user config',
+    path: '$CODEX_HOME/config.toml',
+    resolvedPath: '/tmp/.codex/config.toml',
+    exists: true,
+    isSymlink: false,
+    isRegularFile: true,
+    sizeBytes: 64,
+    mtimeMs: 100,
+    parseError: null,
+    readError: null,
+  },
+  workspacePath: '/tmp/workspace',
+  config: {
+    model: 'gpt-5.3-codex',
+    modelReasoningEffort: null,
+    modelProvider: null,
+    activeProfile: null,
+    approvalPolicy: null,
+    sandboxMode: null,
+    webSearch: null,
+    toolOutputTokenLimit: null,
+    personality: null,
+    topLevelKeys: ['model'],
+    profileCount: 0,
+    profileNames: [],
+    modelProviderCount: 0,
+    modelProviders: [],
+    featureCount: 0,
+    enabledFeatures: [],
+    disabledFeatures: [],
+    trustedProjectCount: 0,
+    untrustedProjectCount: 0,
+    projectTrust: [],
+    mcpServerCount: 0,
+    mcpServers: [],
+  },
+  supportMatrix: [],
+  warnings: [],
+  docsReference: {
+    providerValues: [],
+    settingsHierarchy: [],
+    notes: [],
+    links: [],
+    providerDocs: [],
+  },
+};
+
+const initialRawConfigResponse = {
+  path: '$CODEX_HOME/config.toml',
+  resolvedPath: '/tmp/.codex/config.toml',
+  exists: true,
+  mtime: 100,
+  rawText: 'model = "gpt-5.3-codex"\n',
+  config: { model: 'gpt-5.3-codex' },
+  parseError: null,
+};
+
+const patchedRawConfigResponse = {
+  success: true,
+  path: '$CODEX_HOME/config.toml',
+  resolvedPath: '/tmp/.codex/config.toml',
+  exists: true,
+  mtime: 200,
+  rawText: 'model = "gpt-5.4"\n',
+  config: { model: 'gpt-5.4' },
+  parseError: null,
+};
+
+const wrapper = ({ children }: { children: ReactNode }) => <AllProviders>{children}</AllProviders>;
+
+describe('useCodex', () => {
+  afterEach(() => {
+    vi.unstubAllGlobals();
+    vi.restoreAllMocks();
+  });
+
+  it('updates cached raw config immediately after a structured patch save', async () => {
+    const fetchMock = vi.fn((input: RequestInfo | URL, init?: RequestInit) => {
+      const url = String(input);
+
+      if (url.endsWith('/api/codex/diagnostics')) {
+        return Promise.resolve(createJsonResponse(diagnosticsResponse));
+      }
+
+      if (url.endsWith('/api/codex/config/raw') && !init?.method) {
+        return Promise.resolve(createJsonResponse(initialRawConfigResponse));
+      }
+
+      if (url.endsWith('/api/codex/config/patch')) {
+        return Promise.resolve(createJsonResponse(patchedRawConfigResponse));
+      }
+
+      return Promise.reject(new Error(`Unexpected fetch: ${url}`));
+    });
+
+    vi.stubGlobal('fetch', fetchMock);
+
+    const { result } = renderHook(() => useCodex(), { wrapper });
+
+    await waitFor(() => expect(result.current.rawConfig?.mtime).toBe(100));
+
+    await act(async () => {
+      await result.current.patchConfigAsync({
+        kind: 'top-level',
+        values: { model: 'gpt-5.4' },
+        expectedMtime: 100,
+      });
+    });
+
+    await waitFor(() => expect(result.current.rawConfig?.mtime).toBe(200));
+
+    expect(result.current.rawConfig?.rawText).toBe('model = "gpt-5.4"\n');
+    expect(result.current.rawConfig?.config?.model).toBe('gpt-5.4');
+    expect(
+      fetchMock.mock.calls.filter(([input]) => String(input).endsWith('/api/codex/config/raw'))
+    ).toHaveLength(1);
+  });
+});