Merge remote-tracking branch 'upstream/v0.15.0rc' into refactor/auto-mode-service-gsxdsm

2026-03-16 21:53:07 +00:00 · 2026-02-15 10:20:53 -08:00
parent 8542a32f4f 1662c6bf0b
commit a3a5c9e2cb
42 changed files with 1363 additions and 153 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -98,3 +98,5 @@ data/

 # GSD planning docs (local-only)
 .planning/
+.mcp.json
+.planning
--- a/CLAUDE.md
+++ b/CLAUDE.md
@@ -161,7 +161,7 @@ Use `resolveModelString()` from `@automaker/model-resolver` to convert model ali

 - `haiku` → `claude-haiku-4-5`
 - `sonnet` → `claude-sonnet-4-20250514`
- `opus` → `claude-opus-4-5-20251101`
+- `opus` → `claude-opus-4-6`

 ## Environment Variables

--- a/10
+++ b/10
@@ -118,6 +118,7 @@ RUN curl -fsSL https://opencode.ai/install | bash && \
    echo "=== Checking OpenCode CLI installation ===" && \
    ls -la /home/automaker/.local/bin/ && \
    (which opencode && opencode --version) || echo "opencode installed (may need auth setup)"
+
 USER root

 # Add PATH to profile so it's available in all interactive shells (for login shells)
@@ -147,6 +148,15 @@ COPY --from=server-builder /app/apps/server/package*.json ./apps/server/
 # Copy node_modules (includes symlinks to libs)
 COPY --from=server-builder /app/node_modules ./node_modules

+# Install Playwright Chromium browser for AI agent verification tests
+# This adds ~300MB to the image but enables automated testing mode out of the box
+# Using the locally installed playwright ensures we use the pinned version from package-lock.json
+USER automaker
+RUN ./node_modules/.bin/playwright install chromium && \
+    echo "=== Playwright Chromium installed ===" && \
+    ls -la /home/automaker/.cache/ms-playwright/
+USER root
+
 # Create data and projects directories
 RUN mkdir -p /data /projects && chown automaker:automaker /data /projects

--- a/README.md
+++ b/README.md
@@ -367,6 +367,42 @@ services:

 The Docker image supports both AMD64 and ARM64 architectures. The GitHub CLI and Claude CLI are automatically downloaded for the correct architecture during build.

+##### Playwright for Automated Testing
+
+The Docker image includes **Playwright Chromium pre-installed** for AI agent verification tests. When agents implement features in automated testing mode, they use Playwright to verify the implementation works correctly.
+
+**No additional setup required** - Playwright verification works out of the box.
+
+#### Optional: Persist browsers for manual updates
+
+By default, Playwright Chromium is pre-installed in the Docker image. If you need to manually update browsers or want to persist browser installations across container restarts (not image rebuilds), you can mount a volume.
+
+**Important:** When you first add this volume mount to an existing setup, the empty volume will override the pre-installed browsers. You must re-install them:
+
+```bash
+# After adding the volume mount for the first time
+docker exec --user automaker -w /app automaker-server npx playwright install chromium
+```
+
+Add this to your `docker-compose.override.yml`:
+
+```yaml
+services:
+  server:
+    volumes:
+      - playwright-cache:/home/automaker/.cache/ms-playwright
+
+volumes:
+  playwright-cache:
+    name: automaker-playwright-cache
+```
+
+**Updating browsers manually:**
+
+```bash
+docker exec --user automaker -w /app automaker-server npx playwright install chromium
+```
+
 ### Testing

 #### End-to-End Tests (Playwright)
--- a/apps/server/package.json
+++ b/apps/server/package.json
@@ -24,7 +24,7 @@
    "test:unit": "vitest run tests/unit"
  },
  "dependencies": {
-    "@anthropic-ai/claude-agent-sdk": "0.1.76",
+    "@anthropic-ai/claude-agent-sdk": "0.2.32",
    "@automaker/dependency-resolver": "1.0.0",
    "@automaker/git-utils": "1.0.0",
    "@automaker/model-resolver": "1.0.0",
@@ -34,7 +34,7 @@
    "@automaker/utils": "1.0.0",
    "@github/copilot-sdk": "^0.1.16",
    "@modelcontextprotocol/sdk": "1.25.2",
-    "@openai/codex-sdk": "^0.77.0",
+    "@openai/codex-sdk": "^0.98.0",
    "cookie-parser": "1.4.7",
    "cors": "2.8.5",
    "dotenv": "17.2.3",
@@ -45,6 +45,7 @@
    "yaml": "2.7.0"
  },
  "devDependencies": {
+    "@playwright/test": "1.57.0",
    "@types/cookie": "0.6.0",
    "@types/cookie-parser": "1.4.10",
    "@types/cors": "2.8.19",
--- a/apps/server/src/index.ts
+++ b/apps/server/src/index.ts
@@ -121,21 +121,57 @@ const BOX_CONTENT_WIDTH = 67;
 // The Claude Agent SDK can use either ANTHROPIC_API_KEY or Claude Code CLI authentication
 (async () => {
  const hasAnthropicKey = !!process.env.ANTHROPIC_API_KEY;
+  const hasEnvOAuthToken = !!process.env.CLAUDE_CODE_OAUTH_TOKEN;
+
+  logger.debug('[CREDENTIAL_CHECK] Starting credential detection...');
+  logger.debug('[CREDENTIAL_CHECK] Environment variables:', {
+    hasAnthropicKey,
+    hasEnvOAuthToken,
+  });

  if (hasAnthropicKey) {
    logger.info('✓ ANTHROPIC_API_KEY detected');
    return;
  }

+  if (hasEnvOAuthToken) {
+    logger.info('✓ CLAUDE_CODE_OAUTH_TOKEN detected');
+    return;
+  }
+
  // Check for Claude Code CLI authentication
+  // Store indicators outside the try block so we can use them in the warning message
+  let cliAuthIndicators: Awaited<ReturnType<typeof getClaudeAuthIndicators>> | null = null;
+
  try {
-    const indicators = await getClaudeAuthIndicators();
+    cliAuthIndicators = await getClaudeAuthIndicators();
+    const indicators = cliAuthIndicators;
+
+    // Log detailed credential detection results
+    const { checks, ...indicatorSummary } = indicators;
+    logger.debug('[CREDENTIAL_CHECK] Claude CLI auth indicators:', indicatorSummary);
+
+    logger.debug('[CREDENTIAL_CHECK] File check details:', checks);
+
    const hasCliAuth =
      indicators.hasStatsCacheWithActivity ||
      (indicators.hasSettingsFile && indicators.hasProjectsSessions) ||
      (indicators.hasCredentialsFile &&
        (indicators.credentials?.hasOAuthToken || indicators.credentials?.hasApiKey));

+    logger.debug('[CREDENTIAL_CHECK] Auth determination:', {
+      hasCliAuth,
+      reason: hasCliAuth
+        ? indicators.hasStatsCacheWithActivity
+          ? 'stats cache with activity'
+          : indicators.hasSettingsFile && indicators.hasProjectsSessions
+            ? 'settings file + project sessions'
+            : indicators.credentials?.hasOAuthToken
+              ? 'credentials file with OAuth token'
+              : 'credentials file with API key'
+        : 'no valid credentials found',
+    });
+
    if (hasCliAuth) {
      logger.info('✓ Claude Code CLI authentication detected');
      return;
@@ -145,7 +181,7 @@ const BOX_CONTENT_WIDTH = 67;
    logger.warn('Error checking for Claude Code CLI authentication:', error);
  }

-  // No authentication found - show warning
+  // No authentication found - show warning with paths that were checked
  const wHeader = '⚠️  WARNING: No Claude authentication configured'.padEnd(BOX_CONTENT_WIDTH);
  const w1 = 'The Claude Agent SDK requires authentication to function.'.padEnd(BOX_CONTENT_WIDTH);
  const w2 = 'Options:'.padEnd(BOX_CONTENT_WIDTH);
@@ -158,6 +194,33 @@ const BOX_CONTENT_WIDTH = 67;
    BOX_CONTENT_WIDTH
  );

+  // Build paths checked summary from the indicators (if available)
+  let pathsCheckedInfo = '';
+  if (cliAuthIndicators) {
+    const pathsChecked: string[] = [];
+
+    // Collect paths that were checked (paths are always populated strings)
+    pathsChecked.push(`Settings: ${cliAuthIndicators.checks.settingsFile.path}`);
+    pathsChecked.push(`Stats cache: ${cliAuthIndicators.checks.statsCache.path}`);
+    pathsChecked.push(`Projects dir: ${cliAuthIndicators.checks.projectsDir.path}`);
+    for (const credFile of cliAuthIndicators.checks.credentialFiles) {
+      pathsChecked.push(`Credentials: ${credFile.path}`);
+    }
+
+    if (pathsChecked.length > 0) {
+      pathsCheckedInfo = `
+║                                                                     ║
+║  ${'Paths checked:'.padEnd(BOX_CONTENT_WIDTH)}║
+${pathsChecked
+  .map((p) => {
+    const maxLen = BOX_CONTENT_WIDTH - 4;
+    const display = p.length > maxLen ? '...' + p.slice(-(maxLen - 3)) : p;
+    return `║    ${display.padEnd(maxLen)}  ║`;
+  })
+  .join('\n')}`;
+    }
+  }
+
  logger.warn(`
 ╔═════════════════════════════════════════════════════════════════════╗
 ║  ${wHeader}║
@@ -169,7 +232,7 @@ const BOX_CONTENT_WIDTH = 67;
 ║  ${w3}║
 ║  ${w4}║
 ║  ${w5}║
-║  ${w6}║
+║  ${w6}║${pathsCheckedInfo}
 ║                                                                     ║
 ╚═════════════════════════════════════════════════════════════════════╝
 `);
--- a/apps/server/src/lib/sdk-options.ts
+++ b/apps/server/src/lib/sdk-options.ts
@@ -253,11 +253,27 @@ function buildMcpOptions(config: CreateSdkOptionsConfig): McpOptions {
 /**
 * Build thinking options for SDK configuration.
 * Converts ThinkingLevel to maxThinkingTokens for the Claude SDK.
+ * For adaptive thinking (Opus 4.6), omits maxThinkingTokens to let the model
+ * decide its own reasoning depth.
 *
 * @param thinkingLevel - The thinking level to convert
- * @returns Object with maxThinkingTokens if thinking is enabled
+ * @returns Object with maxThinkingTokens if thinking is enabled with a budget
 */
 function buildThinkingOptions(thinkingLevel?: ThinkingLevel): Partial<Options> {
+  if (!thinkingLevel || thinkingLevel === 'none') {
+    return {};
+  }
+
+  // Adaptive thinking (Opus 4.6): don't set maxThinkingTokens
+  // The model will use adaptive thinking by default
+  if (thinkingLevel === 'adaptive') {
+    logger.debug(
+      `buildThinkingOptions: thinkingLevel="adaptive" -> no maxThinkingTokens (model decides)`
+    );
+    return {};
+  }
+
+  // Manual budget-based thinking for Haiku/Sonnet
  const maxThinkingTokens = getThinkingTokenBudget(thinkingLevel);
  logger.debug(
    `buildThinkingOptions: thinkingLevel="${thinkingLevel}" -> maxThinkingTokens=${maxThinkingTokens}`
--- a/apps/server/src/providers/claude-provider.ts
+++ b/apps/server/src/providers/claude-provider.ts
@@ -219,8 +219,11 @@ export class ClaudeProvider extends BaseProvider {
    // claudeCompatibleProvider takes precedence over claudeApiProfile
    const providerConfig = claudeCompatibleProvider || claudeApiProfile;

-    // Convert thinking level to token budget
-    const maxThinkingTokens = getThinkingTokenBudget(thinkingLevel);
+    // Build thinking configuration
+    // Adaptive thinking (Opus 4.6): don't set maxThinkingTokens, model uses adaptive by default
+    // Manual thinking (Haiku/Sonnet): use budget_tokens
+    const maxThinkingTokens =
+      thinkingLevel === 'adaptive' ? undefined : getThinkingTokenBudget(thinkingLevel);

    // Build Claude SDK options
    const sdkOptions: Options = {
@@ -349,13 +352,13 @@ export class ClaudeProvider extends BaseProvider {
  getAvailableModels(): ModelDefinition[] {
    const models = [
      {
-        id: 'claude-opus-4-5-20251101',
-        name: 'Claude Opus 4.5',
-        modelString: 'claude-opus-4-5-20251101',
+        id: 'claude-opus-4-6',
+        name: 'Claude Opus 4.6',
+        modelString: 'claude-opus-4-6',
        provider: 'anthropic',
-        description: 'Most capable Claude model',
+        description: 'Most capable Claude model with adaptive thinking',
        contextWindow: 200000,
-        maxOutputTokens: 16000,
+        maxOutputTokens: 128000,
        supportsVision: true,
        supportsTools: true,
        tier: 'premium' as const,
--- a/apps/server/src/providers/codex-models.ts
+++ b/apps/server/src/providers/codex-models.ts
@@ -19,12 +19,11 @@ const MAX_OUTPUT_16K = 16000;
 export const CODEX_MODELS: ModelDefinition[] = [
  // ========== Recommended Codex Models ==========
  {
-    id: CODEX_MODEL_MAP.gpt52Codex,
-    name: 'GPT-5.2-Codex',
-    modelString: CODEX_MODEL_MAP.gpt52Codex,
+    id: CODEX_MODEL_MAP.gpt53Codex,
+    name: 'GPT-5.3-Codex',
+    modelString: CODEX_MODEL_MAP.gpt53Codex,
    provider: 'openai',
-    description:
-      'Most advanced agentic coding model for complex software engineering (default for ChatGPT users).',
+    description: 'Latest frontier agentic coding model.',
    contextWindow: CONTEXT_WINDOW_256K,
    maxOutputTokens: MAX_OUTPUT_32K,
    supportsVision: true,
@@ -33,12 +32,25 @@ export const CODEX_MODELS: ModelDefinition[] = [
    default: true,
    hasReasoning: true,
  },
+  {
+    id: CODEX_MODEL_MAP.gpt52Codex,
+    name: 'GPT-5.2-Codex',
+    modelString: CODEX_MODEL_MAP.gpt52Codex,
+    provider: 'openai',
+    description: 'Frontier agentic coding model.',
+    contextWindow: CONTEXT_WINDOW_256K,
+    maxOutputTokens: MAX_OUTPUT_32K,
+    supportsVision: true,
+    supportsTools: true,
+    tier: 'premium' as const,
+    hasReasoning: true,
+  },
  {
    id: CODEX_MODEL_MAP.gpt51CodexMax,
    name: 'GPT-5.1-Codex-Max',
    modelString: CODEX_MODEL_MAP.gpt51CodexMax,
    provider: 'openai',
-    description: 'Optimized for long-horizon, agentic coding tasks in Codex.',
+    description: 'Codex-optimized flagship for deep and fast reasoning.',
    contextWindow: CONTEXT_WINDOW_256K,
    maxOutputTokens: MAX_OUTPUT_32K,
    supportsVision: true,
@@ -51,7 +63,7 @@ export const CODEX_MODELS: ModelDefinition[] = [
    name: 'GPT-5.1-Codex-Mini',
    modelString: CODEX_MODEL_MAP.gpt51CodexMini,
    provider: 'openai',
-    description: 'Smaller, more cost-effective version for faster workflows.',
+    description: 'Optimized for codex. Cheaper, faster, but less capable.',
    contextWindow: CONTEXT_WINDOW_128K,
    maxOutputTokens: MAX_OUTPUT_16K,
    supportsVision: true,
@@ -66,7 +78,7 @@ export const CODEX_MODELS: ModelDefinition[] = [
    name: 'GPT-5.2',
    modelString: CODEX_MODEL_MAP.gpt52,
    provider: 'openai',
-    description: 'Best general agentic model for tasks across industries and domains.',
+    description: 'Latest frontier model with improvements across knowledge, reasoning and coding.',
    contextWindow: CONTEXT_WINDOW_256K,
    maxOutputTokens: MAX_OUTPUT_32K,
    supportsVision: true,
--- a/apps/server/src/providers/provider-factory.ts
+++ b/apps/server/src/providers/provider-factory.ts
@@ -103,7 +103,7 @@ export class ProviderFactory {
  /**
   * Get the appropriate provider for a given model ID
   *
-   * @param modelId Model identifier (e.g., "claude-opus-4-5-20251101", "cursor-gpt-4o", "cursor-auto")
+   * @param modelId Model identifier (e.g., "claude-opus-4-6", "cursor-gpt-4o", "cursor-auto")
   * @param options Optional settings
   * @param options.throwOnDisconnected Throw error if provider is disconnected (default: true)
   * @returns Provider instance for the model
--- a/apps/server/src/routes/setup/routes/verify-claude-auth.ts
+++ b/apps/server/src/routes/setup/routes/verify-claude-auth.ts
@@ -6,6 +6,7 @@
 import type { Request, Response } from 'express';
 import { query } from '@anthropic-ai/claude-agent-sdk';
 import { createLogger } from '@automaker/utils';
+import { getClaudeAuthIndicators } from '@automaker/platform';
 import { getApiKey } from '../common.js';
 import {
  createSecureAuthEnv,
@@ -320,9 +321,28 @@ export function createVerifyClaudeAuthHandler() {
        authMethod,
      });

+      // Determine specific auth type for success messages
+      const effectiveAuthMethod = authMethod ?? 'api_key';
+      let authType: 'oauth' | 'api_key' | 'cli' | undefined;
+      if (authenticated) {
+        if (effectiveAuthMethod === 'api_key') {
+          authType = 'api_key';
+        } else if (effectiveAuthMethod === 'cli') {
+          // Check if CLI auth is via OAuth (Claude Code subscription) or generic CLI
+          try {
+            const indicators = await getClaudeAuthIndicators();
+            authType = indicators.credentials?.hasOAuthToken ? 'oauth' : 'cli';
+          } catch {
+            // Fall back to generic CLI if credential check fails
+            authType = 'cli';
+          }
+        }
+      }
+
      res.json({
        success: true,
        authenticated,
+        authType,
        error: errorMessage || undefined,
      });
    } catch (error) {
--- a/apps/server/tests/unit/lib/model-resolver.test.ts
+++ b/apps/server/tests/unit/lib/model-resolver.test.ts
@@ -35,7 +35,7 @@ describe('model-resolver.ts', () => {

    it("should resolve 'opus' alias to full model string", () => {
      const result = resolveModelString('opus');
-      expect(result).toBe('claude-opus-4-5-20251101');
+      expect(result).toBe('claude-opus-4-6');
      expect(consoleSpy.log).toHaveBeenCalledWith(
        expect.stringContaining('Migrated legacy ID: "opus" -> "claude-opus"')
      );
@@ -117,7 +117,7 @@ describe('model-resolver.ts', () => {
  describe('getEffectiveModel', () => {
    it('should prioritize explicit model over session and default', () => {
      const result = getEffectiveModel('opus', 'haiku', 'gpt-5.2');
-      expect(result).toBe('claude-opus-4-5-20251101');
+      expect(result).toBe('claude-opus-4-6');
    });

    it('should use session model when explicit is not provided', () => {
--- a/apps/server/tests/unit/lib/sdk-options.test.ts
+++ b/apps/server/tests/unit/lib/sdk-options.test.ts
@@ -491,5 +491,29 @@ describe('sdk-options.ts', () => {
        expect(options.maxThinkingTokens).toBeUndefined();
      });
    });
+
+    describe('adaptive thinking for Opus 4.6', () => {
+      it('should not set maxThinkingTokens for adaptive thinking (model decides)', async () => {
+        const { createAutoModeOptions } = await import('@/lib/sdk-options.js');
+
+        const options = createAutoModeOptions({
+          cwd: '/test/path',
+          thinkingLevel: 'adaptive',
+        });
+
+        expect(options.maxThinkingTokens).toBeUndefined();
+      });
+
+      it('should not include maxThinkingTokens when thinkingLevel is "none"', async () => {
+        const { createAutoModeOptions } = await import('@/lib/sdk-options.js');
+
+        const options = createAutoModeOptions({
+          cwd: '/test/path',
+          thinkingLevel: 'none',
+        });
+
+        expect(options.maxThinkingTokens).toBeUndefined();
+      });
+    });
  });
 });
--- a/apps/server/tests/unit/providers/claude-provider.test.ts
+++ b/apps/server/tests/unit/providers/claude-provider.test.ts
@@ -39,7 +39,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Hello',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -59,7 +59,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test prompt',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test/dir',
        systemPrompt: 'You are helpful',
        maxTurns: 10,
@@ -71,7 +71,7 @@ describe('claude-provider.ts', () => {
      expect(sdk.query).toHaveBeenCalledWith({
        prompt: 'Test prompt',
        options: expect.objectContaining({
-          model: 'claude-opus-4-5-20251101',
+          model: 'claude-opus-4-6',
          systemPrompt: 'You are helpful',
          maxTurns: 10,
          cwd: '/test/dir',
@@ -91,7 +91,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -116,7 +116,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
        abortController,
      });
@@ -145,7 +145,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Current message',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
        conversationHistory,
        sdkSessionId: 'test-session-id',
@@ -176,7 +176,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: arrayPrompt as any,
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -196,7 +196,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -222,7 +222,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -286,7 +286,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -313,7 +313,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -341,7 +341,7 @@ describe('claude-provider.ts', () => {

      const generator = provider.executeQuery({
        prompt: 'Test',
-        model: 'claude-opus-4-5-20251101',
+        model: 'claude-opus-4-6',
        cwd: '/test',
      });

@@ -366,12 +366,12 @@ describe('claude-provider.ts', () => {
      expect(models).toHaveLength(4);
    });

-    it('should include Claude Opus 4.5', () => {
+    it('should include Claude Opus 4.6', () => {
      const models = provider.getAvailableModels();

-      const opus = models.find((m) => m.id === 'claude-opus-4-5-20251101');
+      const opus = models.find((m) => m.id === 'claude-opus-4-6');
      expect(opus).toBeDefined();
-      expect(opus?.name).toBe('Claude Opus 4.5');
+      expect(opus?.name).toBe('Claude Opus 4.6');
      expect(opus?.provider).toBe('anthropic');
    });

@@ -400,7 +400,7 @@ describe('claude-provider.ts', () => {
    it('should mark Opus as default', () => {
      const models = provider.getAvailableModels();

-      const opus = models.find((m) => m.id === 'claude-opus-4-5-20251101');
+      const opus = models.find((m) => m.id === 'claude-opus-4-6');
      expect(opus?.default).toBe(true);
    });

--- a/apps/server/tests/unit/providers/provider-factory.test.ts
+++ b/apps/server/tests/unit/providers/provider-factory.test.ts
@@ -54,8 +54,8 @@ describe('provider-factory.ts', () => {

  describe('getProviderForModel', () => {
    describe('Claude models (claude-* prefix)', () => {
-      it('should return ClaudeProvider for claude-opus-4-5-20251101', () => {
-        const provider = ProviderFactory.getProviderForModel('claude-opus-4-5-20251101');
+      it('should return ClaudeProvider for claude-opus-4-6', () => {
+        const provider = ProviderFactory.getProviderForModel('claude-opus-4-6');
        expect(provider).toBeInstanceOf(ClaudeProvider);
      });

@@ -70,7 +70,7 @@ describe('provider-factory.ts', () => {
      });

      it('should be case-insensitive for claude models', () => {
-        const provider = ProviderFactory.getProviderForModel('CLAUDE-OPUS-4-5-20251101');
+        const provider = ProviderFactory.getProviderForModel('CLAUDE-OPUS-4-6');
        expect(provider).toBeInstanceOf(ClaudeProvider);
      });
    });
--- a/apps/ui/docs/AGENT_ARCHITECTURE.md
+++ b/apps/ui/docs/AGENT_ARCHITECTURE.md
@@ -199,7 +199,7 @@ The agent is configured with:

 ```javascript
 {
-  model: "claude-opus-4-5-20251101",
+  model: "claude-opus-4-6",
  maxTurns: 20,
  cwd: workingDirectory,
  allowedTools: [
--- a/apps/ui/src/components/dialogs/sandbox-risk-dialog.tsx
+++ b/apps/ui/src/components/dialogs/sandbox-risk-dialog.tsx
@@ -69,6 +69,29 @@ export function SandboxRiskDialog({ open, onConfirm, onDeny }: SandboxRiskDialog
                For safer operation, consider running Automaker in Docker. See the README for
                instructions.
              </p>
+
+              <div className="bg-muted/50 border border-border rounded-lg p-4 space-y-2">
+                <p className="text-sm font-medium text-foreground">
+                  Already running in Docker? Try these troubleshooting steps:
+                </p>
+                <ul className="text-sm text-muted-foreground list-disc list-inside space-y-1">
+                  <li>
+                    Ensure <code className="bg-muted px-1 rounded">IS_CONTAINERIZED=true</code> is
+                    set in your docker-compose environment
+                  </li>
+                  <li>
+                    Verify the server container has the environment variable:{' '}
+                    <code className="bg-muted px-1 rounded">
+                      docker exec automaker-server printenv IS_CONTAINERIZED
+                    </code>
+                  </li>
+                  <li>Rebuild and restart containers if you recently changed the configuration</li>
+                  <li>
+                    Check the server logs for startup messages:{' '}
+                    <code className="bg-muted px-1 rounded">docker-compose logs server</code>
+                  </li>
+                </ul>
+              </div>
            </div>
          </DialogDescription>
        </DialogHeader>
--- a/apps/ui/src/components/views/board-view/dialogs/add-feature-dialog.tsx
+++ b/apps/ui/src/components/views/board-view/dialogs/add-feature-dialog.tsx
@@ -28,7 +28,7 @@ import { cn } from '@/lib/utils';
 import { modelSupportsThinking } from '@/lib/utils';
 import { useAppStore, ThinkingLevel, FeatureImage, PlanningMode, Feature } from '@/store/app-store';
 import type { ReasoningEffort, PhaseModelEntry, AgentModel } from '@automaker/types';
-import { supportsReasoningEffort } from '@automaker/types';
+import { supportsReasoningEffort, isAdaptiveThinkingModel } from '@automaker/types';
 import {
  PrioritySelector,
  WorkModeSelector,
@@ -264,7 +264,20 @@ export function AddFeatureDialog({
  }, [planningMode]);

  const handleModelChange = (entry: PhaseModelEntry) => {
-    setModelEntry(entry);
+    // Normalize thinking level when switching between adaptive and non-adaptive models
+    const isNewModelAdaptive =
+      typeof entry.model === 'string' && isAdaptiveThinkingModel(entry.model);
+    const currentLevel = entry.thinkingLevel || 'none';
+
+    if (isNewModelAdaptive && currentLevel !== 'none' && currentLevel !== 'adaptive') {
+      // Switching TO Opus 4.6 with a manual level -> auto-switch to 'adaptive'
+      setModelEntry({ ...entry, thinkingLevel: 'adaptive' });
+    } else if (!isNewModelAdaptive && currentLevel === 'adaptive') {
+      // Switching FROM Opus 4.6 with adaptive -> auto-switch to 'high'
+      setModelEntry({ ...entry, thinkingLevel: 'high' });
+    } else {
+      setModelEntry(entry);
+    }
  };

  const buildFeatureData = (): FeatureData | null => {
--- a/apps/ui/src/components/views/board-view/shared/model-constants.ts
+++ b/apps/ui/src/components/views/board-view/shared/model-constants.ts
@@ -167,7 +167,14 @@ export const ALL_MODELS: ModelOption[] = [
  ...COPILOT_MODELS,
 ];

-export const THINKING_LEVELS: ThinkingLevel[] = ['none', 'low', 'medium', 'high', 'ultrathink'];
+export const THINKING_LEVELS: ThinkingLevel[] = [
+  'none',
+  'low',
+  'medium',
+  'high',
+  'ultrathink',
+  'adaptive',
+];

 export const THINKING_LEVEL_LABELS: Record<ThinkingLevel, string> = {
  none: 'None',
@@ -175,6 +182,7 @@ export const THINKING_LEVEL_LABELS: Record<ThinkingLevel, string> = {
  medium: 'Med',
  high: 'High',
  ultrathink: 'Ultra',
+  adaptive: 'Adaptive',
 };

 /**
--- a/apps/ui/src/components/views/board-view/shared/thinking-level-selector.tsx
+++ b/apps/ui/src/components/views/board-view/shared/thinking-level-selector.tsx
@@ -2,19 +2,26 @@ import { Label } from '@/components/ui/label';
 import { Brain } from 'lucide-react';
 import { cn } from '@/lib/utils';
 import { ThinkingLevel } from '@/store/app-store';
-import { THINKING_LEVELS, THINKING_LEVEL_LABELS } from './model-constants';
+import { THINKING_LEVEL_LABELS } from './model-constants';
+import { getThinkingLevelsForModel } from '@automaker/types';

 interface ThinkingLevelSelectorProps {
  selectedLevel: ThinkingLevel;
  onLevelSelect: (level: ThinkingLevel) => void;
  testIdPrefix?: string;
+  /** Model ID is required for correct thinking level filtering.
+   *  Without it, adaptive thinking won't be available for Opus 4.6. */
+  model?: string;
 }

 export function ThinkingLevelSelector({
  selectedLevel,
  onLevelSelect,
  testIdPrefix = 'thinking-level',
+  model,
 }: ThinkingLevelSelectorProps) {
+  const levels = getThinkingLevelsForModel(model || '');
+
  return (
    <div className="space-y-2 pt-2 border-t border-border">
      <Label className="flex items-center gap-2 text-sm">
@@ -22,7 +29,7 @@ export function ThinkingLevelSelector({
        Thinking Level
      </Label>
      <div className="flex gap-2 flex-wrap">
-        {THINKING_LEVELS.map((level) => (
+        {levels.map((level) => (
          <button
            key={level}
            type="button"
@@ -40,7 +47,9 @@ export function ThinkingLevelSelector({
        ))}
      </div>
      <p className="text-xs text-muted-foreground">
-        Higher levels give more time to reason through complex problems.
+        {levels.includes('adaptive')
+          ? 'Adaptive thinking lets the model decide how much reasoning to use.'
+          : 'Higher levels give more time to reason through complex problems.'}
      </p>
    </div>
  );
--- a/apps/ui/src/components/views/settings-view/model-defaults/phase-model-selector.tsx
+++ b/apps/ui/src/components/views/settings-view/model-defaults/phase-model-selector.tsx
@@ -21,6 +21,7 @@ import {
  isGroupSelected,
  getSelectedVariant,
  codexModelHasThinking,
+  getThinkingLevelsForModel,
 } from '@automaker/types';
 import {
  CLAUDE_MODELS,
@@ -28,7 +29,6 @@ import {
  OPENCODE_MODELS,
  GEMINI_MODELS,
  COPILOT_MODELS,
-  THINKING_LEVELS,
  THINKING_LEVEL_LABELS,
  REASONING_EFFORT_LEVELS,
  REASONING_EFFORT_LABELS,
@@ -1296,7 +1296,9 @@ export function PhaseModelSelector({
              <div className="px-2 py-1 text-xs font-medium text-muted-foreground">
                Thinking Level
              </div>
-              {THINKING_LEVELS.map((level) => (
+              {getThinkingLevelsForModel(
+                model.mapsToClaudeModel === 'opus' ? 'claude-opus' : model.id || ''
+              ).map((level) => (
                <button
                  key={level}
                  onClick={() => {
@@ -1322,6 +1324,7 @@ export function PhaseModelSelector({
                      {level === 'medium' && 'Moderate reasoning (10k tokens)'}
                      {level === 'high' && 'Deep reasoning (16k tokens)'}
                      {level === 'ultrathink' && 'Maximum reasoning (32k tokens)'}
+                      {level === 'adaptive' && 'Model decides reasoning depth'}
                    </span>
                  </div>
                  {isSelected && currentThinking === level && (
@@ -1402,7 +1405,9 @@ export function PhaseModelSelector({
              <div className="px-2 py-1.5 text-xs font-medium text-muted-foreground border-b border-border/50 mb-1">
                Thinking Level
              </div>
-              {THINKING_LEVELS.map((level) => (
+              {getThinkingLevelsForModel(
+                model.mapsToClaudeModel === 'opus' ? 'claude-opus' : model.id || ''
+              ).map((level) => (
                <button
                  key={level}
                  onClick={() => {
@@ -1428,6 +1433,7 @@ export function PhaseModelSelector({
                      {level === 'medium' && 'Moderate reasoning (10k tokens)'}
                      {level === 'high' && 'Deep reasoning (16k tokens)'}
                      {level === 'ultrathink' && 'Maximum reasoning (32k tokens)'}
+                      {level === 'adaptive' && 'Model decides reasoning depth'}
                    </span>
                  </div>
                  {isSelected && currentThinking === level && (
@@ -1564,7 +1570,7 @@ export function PhaseModelSelector({
              <div className="px-2 py-1 text-xs font-medium text-muted-foreground">
                Thinking Level
              </div>
-              {THINKING_LEVELS.map((level) => (
+              {getThinkingLevelsForModel(model.id).map((level) => (
                <button
                  key={level}
                  onClick={() => {
@@ -1589,6 +1595,7 @@ export function PhaseModelSelector({
                      {level === 'medium' && 'Moderate reasoning (10k tokens)'}
                      {level === 'high' && 'Deep reasoning (16k tokens)'}
                      {level === 'ultrathink' && 'Maximum reasoning (32k tokens)'}
+                      {level === 'adaptive' && 'Model decides reasoning depth'}
                    </span>
                  </div>
                  {isSelected && currentThinking === level && (
@@ -1685,7 +1692,7 @@ export function PhaseModelSelector({
              <div className="px-2 py-1.5 text-xs font-medium text-muted-foreground border-b border-border/50 mb-1">
                Thinking Level
              </div>
-              {THINKING_LEVELS.map((level) => (
+              {getThinkingLevelsForModel(model.id).map((level) => (
                <button
                  key={level}
                  onClick={() => {
@@ -1710,6 +1717,7 @@ export function PhaseModelSelector({
                      {level === 'medium' && 'Moderate reasoning (10k tokens)'}
                      {level === 'high' && 'Deep reasoning (16k tokens)'}
                      {level === 'ultrathink' && 'Maximum reasoning (32k tokens)'}
+                      {level === 'adaptive' && 'Model decides reasoning depth'}
                    </span>
                  </div>
                  {isSelected && currentThinking === level && (
--- a/apps/ui/src/components/views/settings-view/providers/codex-model-configuration.tsx
+++ b/apps/ui/src/components/views/settings-view/providers/codex-model-configuration.tsx
@@ -9,7 +9,7 @@ import {
  SelectValue,
 } from '@/components/ui/select';
 import { cn } from '@/lib/utils';
-import type { CodexModelId } from '@automaker/types';
+import { supportsReasoningEffort, type CodexModelId } from '@automaker/types';
 import { OpenAIIcon } from '@/components/ui/provider-icon';

 interface CodexModelConfigurationProps {
@@ -27,25 +27,30 @@ interface CodexModelInfo {
 }

 const CODEX_MODEL_INFO: Record<CodexModelId, CodexModelInfo> = {
+  'codex-gpt-5.3-codex': {
+    id: 'codex-gpt-5.3-codex',
+    label: 'GPT-5.3-Codex',
+    description: 'Latest frontier agentic coding model',
+  },
  'codex-gpt-5.2-codex': {
    id: 'codex-gpt-5.2-codex',
    label: 'GPT-5.2-Codex',
-    description: 'Most advanced agentic coding model for complex software engineering',
+    description: 'Frontier agentic coding model',
  },
  'codex-gpt-5.1-codex-max': {
    id: 'codex-gpt-5.1-codex-max',
    label: 'GPT-5.1-Codex-Max',
-    description: 'Optimized for long-horizon, agentic coding tasks in Codex',
+    description: 'Codex-optimized flagship for deep and fast reasoning',
  },
  'codex-gpt-5.1-codex-mini': {
    id: 'codex-gpt-5.1-codex-mini',
    label: 'GPT-5.1-Codex-Mini',
-    description: 'Smaller, more cost-effective version for faster workflows',
+    description: 'Optimized for codex. Cheaper, faster, but less capable',
  },
  'codex-gpt-5.2': {
    id: 'codex-gpt-5.2',
    label: 'GPT-5.2',
-    description: 'Best general agentic model for tasks across industries and domains',
+    description: 'Latest frontier model with improvements across knowledge, reasoning and coding',
  },
  'codex-gpt-5.1': {
    id: 'codex-gpt-5.1',
@@ -157,13 +162,3 @@ export function CodexModelConfiguration({
    </div>
  );
 }
-
-function supportsReasoningEffort(modelId: string): boolean {
-  const reasoningModels = [
-    'codex-gpt-5.2-codex',
-    'codex-gpt-5.1-codex-max',
-    'codex-gpt-5.2',
-    'codex-gpt-5.1',
-  ];
-  return reasoningModels.includes(modelId);
-}
--- a/apps/ui/src/components/views/setup-view/steps/claude-setup-step.tsx
+++ b/apps/ui/src/components/views/setup-view/steps/claude-setup-step.tsx
@@ -59,6 +59,7 @@ export function ClaudeSetupStep({ onNext, onBack, onSkip }: ClaudeSetupStepProps
  // CLI Verification state
  const [cliVerificationStatus, setCliVerificationStatus] = useState<VerificationStatus>('idle');
  const [cliVerificationError, setCliVerificationError] = useState<string | null>(null);
+  const [cliAuthType, setCliAuthType] = useState<'oauth' | 'cli' | null>(null);

  // API Key Verification state
  const [apiKeyVerificationStatus, setApiKeyVerificationStatus] =
@@ -119,6 +120,7 @@ export function ClaudeSetupStep({ onNext, onBack, onSkip }: ClaudeSetupStepProps
  const verifyCliAuth = useCallback(async () => {
    setCliVerificationStatus('verifying');
    setCliVerificationError(null);
+    setCliAuthType(null);

    try {
      const api = getElectronAPI();
@@ -138,12 +140,21 @@ export function ClaudeSetupStep({ onNext, onBack, onSkip }: ClaudeSetupStepProps

      if (result.authenticated && !hasLimitReachedError) {
        setCliVerificationStatus('verified');
+        // Store the auth type for displaying specific success message
+        const authType = result.authType === 'oauth' ? 'oauth' : 'cli';
+        setCliAuthType(authType);
        setClaudeAuthStatus({
          authenticated: true,
-          method: 'cli_authenticated',
+          method: authType === 'oauth' ? 'oauth_token' : 'cli_authenticated',
          hasCredentialsFile: claudeAuthStatus?.hasCredentialsFile || false,
+          oauthTokenValid: authType === 'oauth',
        });
-        toast.success('Claude CLI authentication verified!');
+        // Show specific success message based on auth type
+        if (authType === 'oauth') {
+          toast.success('Claude Code subscription detected and verified!');
+        } else {
+          toast.success('Claude CLI authentication verified!');
+        }
      } else {
        setCliVerificationStatus('error');
        setCliVerificationError(
@@ -436,9 +447,15 @@ export function ClaudeSetupStep({ onNext, onBack, onSkip }: ClaudeSetupStepProps
                  <div className="flex items-center gap-3 p-4 rounded-lg bg-green-500/10 border border-green-500/20">
                    <CheckCircle2 className="w-5 h-5 text-green-500" />
                    <div>
-                      <p className="font-medium text-foreground">CLI Authentication verified!</p>
+                      <p className="font-medium text-foreground">
+                        {cliAuthType === 'oauth'
+                          ? 'Claude Code subscription verified!'
+                          : 'CLI Authentication verified!'}
+                      </p>
                      <p className="text-sm text-muted-foreground">
-                        Your Claude CLI is working correctly.
+                        {cliAuthType === 'oauth'
+                          ? 'Your Claude Code subscription is active and ready to use.'
+                          : 'Your Claude CLI is working correctly.'}
                      </p>
                    </div>
                  </div>
--- a/apps/ui/src/lib/agent-context-parser.ts
+++ b/apps/ui/src/lib/agent-context-parser.ts
@@ -27,18 +27,20 @@ export interface AgentTaskInfo {
 /**
 * Default model used by the feature executor
 */
-export const DEFAULT_MODEL = 'claude-opus-4-5-20251101';
+export const DEFAULT_MODEL = 'claude-opus-4-6';

 /**
 * Formats a model name for display
 */
 export function formatModelName(model: string): string {
  // Claude models
+  if (model.includes('opus-4-6') || model === 'claude-opus') return 'Opus 4.6';
  if (model.includes('opus')) return 'Opus 4.5';
  if (model.includes('sonnet')) return 'Sonnet 4.5';
  if (model.includes('haiku')) return 'Haiku 4.5';

  // Codex/GPT models - specific formatting
+  if (model === 'codex-gpt-5.3-codex') return 'GPT-5.3 Codex';
  if (model === 'codex-gpt-5.2-codex') return 'GPT-5.2 Codex';
  if (model === 'codex-gpt-5.2') return 'GPT-5.2';
  if (model === 'codex-gpt-5.1-codex-max') return 'GPT-5.1 Max';
--- a/apps/ui/src/lib/electron.ts
+++ b/apps/ui/src/lib/electron.ts
@@ -1442,6 +1442,7 @@ interface SetupAPI {
  verifyClaudeAuth: (authMethod?: 'cli' | 'api_key') => Promise<{
    success: boolean;
    authenticated: boolean;
+    authType?: 'oauth' | 'api_key' | 'cli';
    error?: string;
  }>;
  getGhStatus?: () => Promise<{
--- a/apps/ui/src/lib/http-api-client.ts
+++ b/apps/ui/src/lib/http-api-client.ts
@@ -1350,6 +1350,7 @@ export class HttpApiClient implements ElectronAPI {
    ): Promise<{
      success: boolean;
      authenticated: boolean;
+      authType?: 'oauth' | 'api_key' | 'cli';
      error?: string;
    }> => this.post('/api/setup/verify-claude-auth', { authMethod, apiKey }),

--- a/docker-compose.override.yml.example
+++ b/docker-compose.override.yml.example
@@ -21,9 +21,13 @@ services:
      # - ~/.local/share/opencode:/home/automaker/.local/share/opencode
      # - ~/.config/opencode:/home/automaker/.config/opencode

-      # Playwright browser cache - persists installed browsers across container restarts
-      # Run 'npx playwright install --with-deps chromium' once, and it will persist
+      # ===== Playwright Browser Cache (Optional) =====
+      # Playwright Chromium is PRE-INSTALLED in the Docker image for automated testing.
+      # Uncomment below to persist browser cache across container rebuilds (saves ~300MB download):
      # - playwright-cache:/home/automaker/.cache/ms-playwright
+      #
+      # To update Playwright browsers manually:
+      #   docker exec --user automaker -w /app automaker-server npx playwright install chromium
    environment:
      # Set root directory for all projects and file operations
      # Users can only create/open projects within this directory
@@ -37,6 +41,7 @@ services:
      # - CURSOR_API_KEY=${CURSOR_API_KEY:-}

 volumes:
-  # Playwright cache volume (persists Chromium installs)
+  # Playwright cache volume - optional, persists browser updates across container rebuilds
+  # Uncomment if you mounted the playwright-cache volume above
  # playwright-cache:
  #   name: automaker-playwright-cache
--- a/docs/llm-shared-packages.md
+++ b/docs/llm-shared-packages.md
@@ -142,7 +142,7 @@ const modelId = resolveModelString('sonnet'); // → 'claude-sonnet-4-20250514'

 - `haiku` → `claude-haiku-4-5` (fast, simple tasks)
 - `sonnet` → `claude-sonnet-4-20250514` (balanced, recommended)
- `opus` → `claude-opus-4-5-20251101` (maximum capability)
+- `opus` → `claude-opus-4-6` (maximum capability)

 ### @automaker/dependency-resolver

--- a/docs/server/providers.md
+++ b/docs/server/providers.md
@@ -175,7 +175,7 @@ Uses `@anthropic-ai/claude-agent-sdk` for direct SDK integration.

 Routes models that:

- Start with `"claude-"` (e.g., `"claude-opus-4-5-20251101"`)
+- Start with `"claude-"` (e.g., `"claude-opus-4-6"`)
 - Are Claude aliases: `"opus"`, `"sonnet"`, `"haiku"`

 #### Authentication
@@ -191,7 +191,7 @@ const provider = new ClaudeProvider();

 const stream = provider.executeQuery({
  prompt: 'What is 2+2?',
-  model: 'claude-opus-4-5-20251101',
+  model: 'claude-opus-4-6',
  cwd: '/project/path',
  systemPrompt: 'You are a helpful assistant.',
  maxTurns: 20,
@@ -701,7 +701,7 @@ Test provider interaction with services:
 ```typescript
 describe('Provider Integration', () => {
  it('should work with AgentService', async () => {
-    const provider = ProviderFactory.getProviderForModel('claude-opus-4-5-20251101');
+    const provider = ProviderFactory.getProviderForModel('claude-opus-4-6');

    // Test full workflow
  });
--- a/docs/server/utilities.md
+++ b/docs/server/utilities.md
@@ -213,7 +213,7 @@ Model alias mapping for Claude models.
 export const CLAUDE_MODEL_MAP: Record<string, string> = {
  haiku: 'claude-haiku-4-5',
  sonnet: 'claude-sonnet-4-20250514',
-  opus: 'claude-opus-4-5-20251101',
+  opus: 'claude-opus-4-6',
 } as const;
 ```

@@ -223,7 +223,7 @@ Default models per provider.

 ```typescript
 export const DEFAULT_MODELS = {
-  claude: 'claude-opus-4-5-20251101',
+  claude: 'claude-opus-4-6',
  openai: 'gpt-5.2',
 } as const;
 ```
@@ -248,8 +248,8 @@ Resolve a model key/alias to a full model string.
 import { resolveModelString, DEFAULT_MODELS } from '../lib/model-resolver.js';

 resolveModelString('opus');
-// Returns: "claude-opus-4-5-20251101"
-// Logs: "[ModelResolver] Resolved model alias: "opus" -> "claude-opus-4-5-20251101""
+// Returns: "claude-opus-4-6"
+// Logs: "[ModelResolver] Resolved model alias: "opus" -> "claude-opus-4-6""

 resolveModelString('gpt-5.2');
 // Returns: "gpt-5.2"
@@ -260,8 +260,8 @@ resolveModelString('claude-sonnet-4-20250514');
 // Logs: "[ModelResolver] Using full Claude model string: claude-sonnet-4-20250514"

 resolveModelString('invalid-model');
-// Returns: "claude-opus-4-5-20251101"
-// Logs: "[ModelResolver] Unknown model key "invalid-model", using default: "claude-opus-4-5-20251101""
+// Returns: "claude-opus-4-6"
+// Logs: "[ModelResolver] Unknown model key "invalid-model", using default: "claude-opus-4-6""
 ```

 ---
--- a/libs/model-resolver/README.md
+++ b/libs/model-resolver/README.md
@@ -30,15 +30,15 @@ const model2 = resolveModelString('haiku');
 // Returns: 'claude-haiku-4-5'

 const model3 = resolveModelString('opus');
-// Returns: 'claude-opus-4-5-20251101'
+// Returns: 'claude-opus-4-6'

 // Use with custom default
 const model4 = resolveModelString(undefined, 'claude-sonnet-4-20250514');
 // Returns: 'claude-sonnet-4-20250514' (default)

 // Direct model ID passthrough
-const model5 = resolveModelString('claude-opus-4-5-20251101');
-// Returns: 'claude-opus-4-5-20251101' (unchanged)
+const model5 = resolveModelString('claude-opus-4-6');
+// Returns: 'claude-opus-4-6' (unchanged)
 ```

 ### Get Effective Model
@@ -72,7 +72,7 @@ console.log(DEFAULT_MODELS.chat); // 'claude-sonnet-4-20250514'
 // Model alias mappings
 console.log(CLAUDE_MODEL_MAP.haiku); // 'claude-haiku-4-5'
 console.log(CLAUDE_MODEL_MAP.sonnet); // 'claude-sonnet-4-20250514'
-console.log(CLAUDE_MODEL_MAP.opus); // 'claude-opus-4-5-20251101'
+console.log(CLAUDE_MODEL_MAP.opus); // 'claude-opus-4-6'
 ```

 ## Usage Example
@@ -103,7 +103,7 @@ const feature: Feature = {
 };

 prepareFeatureExecution(feature);
-// Output: Executing feature with model: claude-opus-4-5-20251101
+// Output: Executing feature with model: claude-opus-4-6
 ```

 ## Supported Models
@@ -112,7 +112,7 @@ prepareFeatureExecution(feature);

 - `haiku` → `claude-haiku-4-5`
 - `sonnet` → `claude-sonnet-4-20250514`
- `opus` → `claude-opus-4-5-20251101`
+- `opus` → `claude-opus-4-6`

 ### Model Selection Guide

--- a/libs/model-resolver/tests/resolver.test.ts
+++ b/libs/model-resolver/tests/resolver.test.ts
@@ -484,12 +484,12 @@ describe('model-resolver', () => {

      it('should handle full Claude model string in entry', () => {
        const entry: PhaseModelEntry = {
-          model: 'claude-opus-4-5-20251101',
+          model: 'claude-opus-4-6',
          thinkingLevel: 'high',
        };
        const result = resolvePhaseModel(entry);

-        expect(result.model).toBe('claude-opus-4-5-20251101');
+        expect(result.model).toBe('claude-opus-4-6');
        expect(result.thinkingLevel).toBe('high');
      });
    });
--- a/libs/platform/src/index.ts
+++ b/libs/platform/src/index.ts
@@ -134,6 +134,8 @@ export {
  findClaudeCliPath,
  getClaudeAuthIndicators,
  type ClaudeAuthIndicators,
+  type FileCheckResult,
+  type DirectoryCheckResult,
  findCodexCliPath,
  getCodexAuthIndicators,
  type CodexAuthIndicators,
--- a/libs/platform/src/system-paths.ts
+++ b/libs/platform/src/system-paths.ts
@@ -25,6 +25,16 @@ import fs from 'fs/promises';
 // System Tool Path Definitions
 // =============================================================================

+/**
+ * Get NVM for Windows (nvm4w) symlink paths for a given CLI tool.
+ * Reused across getClaudeCliPaths, getCodexCliPaths, and getOpenCodeCliPaths.
+ */
+function getNvmWindowsCliPaths(cliName: string): string[] {
+  const nvmSymlink = process.env.NVM_SYMLINK;
+  if (!nvmSymlink) return [];
+  return [path.join(nvmSymlink, `${cliName}.cmd`), path.join(nvmSymlink, cliName)];
+}
+
 /**
 * Get common paths where GitHub CLI might be installed
 */
@@ -60,6 +70,7 @@ export function getClaudeCliPaths(): string[] {
      path.join(appData, 'npm', 'claude'),
      path.join(appData, '.npm-global', 'bin', 'claude.cmd'),
      path.join(appData, '.npm-global', 'bin', 'claude'),
+      ...getNvmWindowsCliPaths('claude'),
    ];
  }

@@ -141,6 +152,7 @@ export function getCodexCliPaths(): string[] {
      // pnpm on Windows
      path.join(localAppData, 'pnpm', 'codex.cmd'),
      path.join(localAppData, 'pnpm', 'codex'),
+      ...getNvmWindowsCliPaths('codex'),
    ];
  }

@@ -976,6 +988,27 @@ export async function findGitBashPath(): Promise<string | null> {
  return findFirstExistingPath(getGitBashPaths());
 }

+/**
+ * Details about a file check performed during auth detection
+ */
+export interface FileCheckResult {
+  path: string;
+  exists: boolean;
+  readable: boolean;
+  error?: string;
+}
+
+/**
+ * Details about a directory check performed during auth detection
+ */
+export interface DirectoryCheckResult {
+  path: string;
+  exists: boolean;
+  readable: boolean;
+  entryCount: number;
+  error?: string;
+}
+
 /**
 * Get Claude authentication status by checking various indicators
 */
@@ -988,67 +1021,165 @@ export interface ClaudeAuthIndicators {
    hasOAuthToken: boolean;
    hasApiKey: boolean;
  } | null;
+  /** Detailed information about what was checked */
+  checks: {
+    settingsFile: FileCheckResult;
+    statsCache: FileCheckResult & { hasDailyActivity?: boolean };
+    projectsDir: DirectoryCheckResult;
+    credentialFiles: FileCheckResult[];
+  };
 }

 export async function getClaudeAuthIndicators(): Promise<ClaudeAuthIndicators> {
+  const settingsPath = getClaudeSettingsPath();
+  const statsCachePath = getClaudeStatsCachePath();
+  const projectsDir = getClaudeProjectsDir();
+  const credentialPaths = getClaudeCredentialPaths();
+
+  // Initialize checks with paths
+  const settingsFileCheck: FileCheckResult = {
+    path: settingsPath,
+    exists: false,
+    readable: false,
+  };
+
+  const statsCacheCheck: FileCheckResult & { hasDailyActivity?: boolean } = {
+    path: statsCachePath,
+    exists: false,
+    readable: false,
+  };
+
+  const projectsDirCheck: DirectoryCheckResult = {
+    path: projectsDir,
+    exists: false,
+    readable: false,
+    entryCount: 0,
+  };
+
+  const credentialFileChecks: FileCheckResult[] = credentialPaths.map((p) => ({
+    path: p,
+    exists: false,
+    readable: false,
+  }));
+
  const result: ClaudeAuthIndicators = {
    hasCredentialsFile: false,
    hasSettingsFile: false,
    hasStatsCacheWithActivity: false,
    hasProjectsSessions: false,
    credentials: null,
+    checks: {
+      settingsFile: settingsFileCheck,
+      statsCache: statsCacheCheck,
+      projectsDir: projectsDirCheck,
+      credentialFiles: credentialFileChecks,
+    },
  };

  // Check settings file
+  // First check existence, then try to read to confirm it's actually readable
  try {
-    if (await systemPathAccess(getClaudeSettingsPath())) {
-      result.hasSettingsFile = true;
+    if (await systemPathAccess(settingsPath)) {
+      settingsFileCheck.exists = true;
+      // Try to actually read the file to confirm read permissions
+      try {
+        await systemPathReadFile(settingsPath);
+        settingsFileCheck.readable = true;
+        result.hasSettingsFile = true;
+      } catch (readErr) {
+        // File exists but cannot be read (permission denied, etc.)
+        settingsFileCheck.readable = false;
+        settingsFileCheck.error = `Cannot read: ${readErr instanceof Error ? readErr.message : String(readErr)}`;
+      }
    }
-  } catch {
-    // Ignore errors
+  } catch (err) {
+    settingsFileCheck.error = err instanceof Error ? err.message : String(err);
  }

  // Check stats cache for recent activity
  try {
-    const statsContent = await systemPathReadFile(getClaudeStatsCachePath());
-    const stats = JSON.parse(statsContent);
-    if (stats.dailyActivity && stats.dailyActivity.length > 0) {
-      result.hasStatsCacheWithActivity = true;
+    const statsContent = await systemPathReadFile(statsCachePath);
+    statsCacheCheck.exists = true;
+    statsCacheCheck.readable = true;
+    try {
+      const stats = JSON.parse(statsContent);
+      if (stats.dailyActivity && stats.dailyActivity.length > 0) {
+        statsCacheCheck.hasDailyActivity = true;
+        result.hasStatsCacheWithActivity = true;
+      } else {
+        statsCacheCheck.hasDailyActivity = false;
+      }
+    } catch (parseErr) {
+      statsCacheCheck.error = `JSON parse error: ${parseErr instanceof Error ? parseErr.message : String(parseErr)}`;
+    }
+  } catch (err) {
+    if ((err as NodeJS.ErrnoException).code === 'ENOENT') {
+      statsCacheCheck.exists = false;
+    } else {
+      statsCacheCheck.error = err instanceof Error ? err.message : String(err);
    }
-  } catch {
-    // Ignore errors
  }

  // Check for sessions in projects directory
  try {
-    const sessions = await systemPathReaddir(getClaudeProjectsDir());
+    const sessions = await systemPathReaddir(projectsDir);
+    projectsDirCheck.exists = true;
+    projectsDirCheck.readable = true;
+    projectsDirCheck.entryCount = sessions.length;
    if (sessions.length > 0) {
      result.hasProjectsSessions = true;
    }
-  } catch {
-    // Ignore errors
+  } catch (err) {
+    if ((err as NodeJS.ErrnoException).code === 'ENOENT') {
+      projectsDirCheck.exists = false;
+    } else {
+      projectsDirCheck.error = err instanceof Error ? err.message : String(err);
+    }
  }

  // Check credentials files
-  const credentialPaths = getClaudeCredentialPaths();
-  for (const credPath of credentialPaths) {
+  // We iterate through all credential paths and only stop when we find a file
+  // that contains actual credentials (OAuth tokens or API keys). An empty or
+  // token-less file should not prevent checking subsequent credential paths.
+  for (let i = 0; i < credentialPaths.length; i++) {
+    const credPath = credentialPaths[i];
+    const credCheck = credentialFileChecks[i];
    try {
      const content = await systemPathReadFile(credPath);
-      const credentials = JSON.parse(content);
-      result.hasCredentialsFile = true;
-      // Support multiple credential formats:
-      // 1. Claude Code CLI format: { claudeAiOauth: { accessToken, refreshToken } }
-      // 2. Legacy format: { oauth_token } or { access_token }
-      // 3. API key format: { api_key }
-      const hasClaudeOauth = !!credentials.claudeAiOauth?.accessToken;
-      const hasLegacyOauth = !!(credentials.oauth_token || credentials.access_token);
-      result.credentials = {
-        hasOAuthToken: hasClaudeOauth || hasLegacyOauth,
-        hasApiKey: !!credentials.api_key,
-      };
-      break;
-    } catch {
-      // Continue to next path
+      credCheck.exists = true;
+      credCheck.readable = true;
+      try {
+        const credentials = JSON.parse(content);
+        // Support multiple credential formats:
+        // 1. Claude Code CLI format: { claudeAiOauth: { accessToken, refreshToken } }
+        // 2. Legacy format: { oauth_token } or { access_token }
+        // 3. API key format: { api_key }
+        const hasClaudeOauth = !!credentials.claudeAiOauth?.accessToken;
+        const hasLegacyOauth = !!(credentials.oauth_token || credentials.access_token);
+        const hasOAuthToken = hasClaudeOauth || hasLegacyOauth;
+        const hasApiKey = !!credentials.api_key;
+
+        // Only consider this a valid credentials file if it actually contains tokens
+        // An empty JSON file ({}) or file without tokens should not stop us from
+        // checking subsequent credential paths
+        if (hasOAuthToken || hasApiKey) {
+          result.hasCredentialsFile = true;
+          result.credentials = {
+            hasOAuthToken,
+            hasApiKey,
+          };
+          break; // Found valid credentials, stop searching
+        }
+        // File exists and is valid JSON but contains no tokens - continue checking other paths
+      } catch (parseErr) {
+        credCheck.error = `JSON parse error: ${parseErr instanceof Error ? parseErr.message : String(parseErr)}`;
+      }
+    } catch (err) {
+      if ((err as NodeJS.ErrnoException).code === 'ENOENT') {
+        credCheck.exists = false;
+      } else {
+        credCheck.error = err instanceof Error ? err.message : String(err);
+      }
    }
  }

@@ -1142,6 +1273,7 @@ export function getOpenCodeCliPaths(): string[] {
      // Go installation (if OpenCode is a Go binary)
      path.join(homeDir, 'go', 'bin', 'opencode.exe'),
      path.join(process.env.GOPATH || path.join(homeDir, 'go'), 'bin', 'opencode.exe'),
+      ...getNvmWindowsCliPaths('opencode'),
    ];
  }

--- a/libs/platform/tests/oauth-credential-detection.test.ts
+++ b/libs/platform/tests/oauth-credential-detection.test.ts
@@ -0,0 +1,761 @@
+/**
+ * Unit tests for OAuth credential detection scenarios
+ *
+ * Tests the various Claude credential detection formats including:
+ * - Claude Code CLI OAuth format (claudeAiOauth)
+ * - Legacy OAuth token format (oauth_token, access_token)
+ * - API key format (api_key)
+ * - Invalid/malformed credential files
+ *
+ * These tests use real temp directories to avoid complex fs mocking issues.
+ */
+
+import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest';
+import fs from 'fs/promises';
+import path from 'path';
+import os from 'os';
+
+describe('OAuth Credential Detection', () => {
+  let tempDir: string;
+  let originalHomedir: () => string;
+  let mockClaudeDir: string;
+  let mockCodexDir: string;
+  let mockOpenCodeDir: string;
+
+  beforeEach(async () => {
+    // Reset modules to get fresh state
+    vi.resetModules();
+
+    // Create a temporary directory
+    tempDir = await fs.mkdtemp(path.join(os.tmpdir(), 'oauth-detection-test-'));
+
+    // Create mock home directory structure
+    mockClaudeDir = path.join(tempDir, '.claude');
+    mockCodexDir = path.join(tempDir, '.codex');
+    mockOpenCodeDir = path.join(tempDir, '.local', 'share', 'opencode');
+
+    await fs.mkdir(mockClaudeDir, { recursive: true });
+    await fs.mkdir(mockCodexDir, { recursive: true });
+    await fs.mkdir(mockOpenCodeDir, { recursive: true });
+
+    // Mock os.homedir to return our temp directory
+    originalHomedir = os.homedir;
+    vi.spyOn(os, 'homedir').mockReturnValue(tempDir);
+  });
+
+  afterEach(async () => {
+    vi.restoreAllMocks();
+    // Clean up temp directory
+    try {
+      await fs.rm(tempDir, { recursive: true, force: true });
+    } catch {
+      // Ignore cleanup errors
+    }
+  });
+
+  describe('getClaudeAuthIndicators', () => {
+    it('should detect Claude Code CLI OAuth format (claudeAiOauth)', async () => {
+      const credentialsContent = JSON.stringify({
+        claudeAiOauth: {
+          accessToken: 'oauth-access-token-12345',
+          refreshToken: 'oauth-refresh-token-67890',
+          expiresAt: Date.now() + 3600000,
+        },
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), credentialsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials).not.toBeNull();
+      expect(indicators.credentials?.hasOAuthToken).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(false);
+    });
+
+    it('should detect legacy OAuth token format (oauth_token)', async () => {
+      const credentialsContent = JSON.stringify({
+        oauth_token: 'legacy-oauth-token-abcdef',
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), credentialsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasOAuthToken).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(false);
+    });
+
+    it('should detect legacy access_token format', async () => {
+      const credentialsContent = JSON.stringify({
+        access_token: 'legacy-access-token-xyz',
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), credentialsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasOAuthToken).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(false);
+    });
+
+    it('should detect API key format', async () => {
+      const credentialsContent = JSON.stringify({
+        api_key: 'sk-ant-api03-xxxxxxxxxxxx',
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), credentialsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasOAuthToken).toBe(false);
+      expect(indicators.credentials?.hasApiKey).toBe(true);
+    });
+
+    it('should detect both OAuth and API key when present', async () => {
+      const credentialsContent = JSON.stringify({
+        claudeAiOauth: {
+          accessToken: 'oauth-token',
+          refreshToken: 'refresh-token',
+        },
+        api_key: 'sk-ant-api03-xxxxxxxxxxxx',
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), credentialsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasOAuthToken).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(true);
+    });
+
+    it('should handle missing credentials file gracefully', async () => {
+      // No credentials file created
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasCredentialsFile).toBe(false);
+      expect(indicators.credentials).toBeNull();
+      expect(indicators.checks.credentialFiles).toBeDefined();
+      expect(indicators.checks.credentialFiles.length).toBeGreaterThan(0);
+      expect(indicators.checks.credentialFiles[0].exists).toBe(false);
+    });
+
+    it('should handle malformed JSON in credentials file', async () => {
+      const malformedContent = '{ invalid json }';
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), malformedContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // File exists but parsing fails
+      expect(indicators.hasCredentialsFile).toBe(false);
+      expect(indicators.credentials).toBeNull();
+      expect(indicators.checks.credentialFiles[0].exists).toBe(true);
+      expect(indicators.checks.credentialFiles[0].error).toContain('JSON parse error');
+    });
+
+    it('should handle empty credentials file', async () => {
+      const emptyContent = JSON.stringify({});
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), emptyContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Empty credentials file ({}) should NOT be treated as having credentials
+      // because it contains no actual tokens. This allows the system to continue
+      // checking subsequent credential paths that might have valid tokens.
+      expect(indicators.hasCredentialsFile).toBe(false);
+      expect(indicators.credentials).toBeNull();
+      // But the file should still show as existing and readable in the checks
+      expect(indicators.checks.credentialFiles[0].exists).toBe(true);
+      expect(indicators.checks.credentialFiles[0].readable).toBe(true);
+    });
+
+    it('should handle credentials file with null values', async () => {
+      const nullContent = JSON.stringify({
+        claudeAiOauth: null,
+        api_key: null,
+        oauth_token: null,
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), nullContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // File with all null values should NOT be treated as having credentials
+      // because null values are not valid tokens
+      expect(indicators.hasCredentialsFile).toBe(false);
+      expect(indicators.credentials).toBeNull();
+    });
+
+    it('should handle credentials with empty string values', async () => {
+      const emptyStrings = JSON.stringify({
+        claudeAiOauth: {
+          accessToken: '',
+          refreshToken: '',
+        },
+        api_key: '',
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), emptyStrings);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Empty strings should NOT be treated as having credentials
+      // This allows checking subsequent credential paths for valid tokens
+      expect(indicators.hasCredentialsFile).toBe(false);
+      expect(indicators.credentials).toBeNull();
+    });
+
+    it('should detect settings file presence', async () => {
+      await fs.writeFile(
+        path.join(mockClaudeDir, 'settings.json'),
+        JSON.stringify({ theme: 'dark' })
+      );
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasSettingsFile).toBe(true);
+      expect(indicators.checks.settingsFile.exists).toBe(true);
+      expect(indicators.checks.settingsFile.readable).toBe(true);
+    });
+
+    it('should detect stats cache with activity', async () => {
+      const statsContent = JSON.stringify({
+        dailyActivity: [
+          { date: '2025-01-15', messagesCount: 10 },
+          { date: '2025-01-16', messagesCount: 5 },
+        ],
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, 'stats-cache.json'), statsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasStatsCacheWithActivity).toBe(true);
+      expect(indicators.checks.statsCache.exists).toBe(true);
+      expect(indicators.checks.statsCache.hasDailyActivity).toBe(true);
+    });
+
+    it('should detect stats cache without activity', async () => {
+      const statsContent = JSON.stringify({
+        dailyActivity: [],
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, 'stats-cache.json'), statsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasStatsCacheWithActivity).toBe(false);
+      expect(indicators.checks.statsCache.exists).toBe(true);
+      expect(indicators.checks.statsCache.hasDailyActivity).toBe(false);
+    });
+
+    it('should detect project sessions', async () => {
+      const projectsDir = path.join(mockClaudeDir, 'projects');
+      await fs.mkdir(projectsDir, { recursive: true });
+      await fs.mkdir(path.join(projectsDir, 'session-1'));
+      await fs.mkdir(path.join(projectsDir, 'session-2'));
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasProjectsSessions).toBe(true);
+      expect(indicators.checks.projectsDir.exists).toBe(true);
+      expect(indicators.checks.projectsDir.entryCount).toBe(2);
+    });
+
+    it('should return comprehensive check details', async () => {
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Verify all check detail objects are present
+      expect(indicators.checks).toBeDefined();
+      expect(indicators.checks.settingsFile).toBeDefined();
+      expect(indicators.checks.settingsFile.path).toContain('settings.json');
+      expect(indicators.checks.statsCache).toBeDefined();
+      expect(indicators.checks.statsCache.path).toContain('stats-cache.json');
+      expect(indicators.checks.projectsDir).toBeDefined();
+      expect(indicators.checks.projectsDir.path).toContain('projects');
+      expect(indicators.checks.credentialFiles).toBeDefined();
+      expect(Array.isArray(indicators.checks.credentialFiles)).toBe(true);
+    });
+
+    it('should try both .credentials.json and credentials.json paths', async () => {
+      // Write to credentials.json (without leading dot)
+      const credentialsContent = JSON.stringify({
+        api_key: 'sk-test-key',
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, 'credentials.json'), credentialsContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Should find credentials in the second path
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(true);
+    });
+
+    it('should prefer first credentials file if both exist', async () => {
+      // Write OAuth to .credentials.json (first path checked)
+      await fs.writeFile(
+        path.join(mockClaudeDir, '.credentials.json'),
+        JSON.stringify({
+          claudeAiOauth: {
+            accessToken: 'oauth-token',
+            refreshToken: 'refresh-token',
+          },
+        })
+      );
+
+      // Write API key to credentials.json (second path)
+      await fs.writeFile(
+        path.join(mockClaudeDir, 'credentials.json'),
+        JSON.stringify({
+          api_key: 'sk-test-key',
+        })
+      );
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Should use first file (.credentials.json) which has OAuth
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasOAuthToken).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(false);
+    });
+
+    it('should check second credentials file if first file has no tokens', async () => {
+      // Write empty/token-less content to .credentials.json (first path checked)
+      // This tests the bug fix: previously, an empty JSON file would stop the search
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), JSON.stringify({}));
+
+      // Write actual credentials to credentials.json (second path)
+      await fs.writeFile(
+        path.join(mockClaudeDir, 'credentials.json'),
+        JSON.stringify({
+          api_key: 'sk-test-key-from-second-file',
+        })
+      );
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Should find credentials in second file since first file has no tokens
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(true);
+    });
+  });
+
+  describe('getCodexAuthIndicators', () => {
+    it('should detect OAuth token in Codex auth file', async () => {
+      const authContent = JSON.stringify({
+        access_token: 'codex-oauth-token-12345',
+      });
+
+      await fs.writeFile(path.join(mockCodexDir, 'auth.json'), authContent);
+
+      const { getCodexAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getCodexAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(true);
+      expect(indicators.hasApiKey).toBe(false);
+    });
+
+    it('should detect API key in Codex auth file', async () => {
+      const authContent = JSON.stringify({
+        OPENAI_API_KEY: 'sk-xxxxxxxxxxxxxxxx',
+      });
+
+      await fs.writeFile(path.join(mockCodexDir, 'auth.json'), authContent);
+
+      const { getCodexAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getCodexAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(false);
+      expect(indicators.hasApiKey).toBe(true);
+    });
+
+    it('should detect nested tokens in Codex auth file', async () => {
+      const authContent = JSON.stringify({
+        tokens: {
+          oauth_token: 'nested-oauth-token',
+        },
+      });
+
+      await fs.writeFile(path.join(mockCodexDir, 'auth.json'), authContent);
+
+      const { getCodexAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getCodexAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(true);
+    });
+
+    it('should handle missing Codex auth file', async () => {
+      // No auth file created
+      const { getCodexAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getCodexAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(false);
+      expect(indicators.hasOAuthToken).toBe(false);
+      expect(indicators.hasApiKey).toBe(false);
+    });
+
+    it('should detect api_key field in Codex auth', async () => {
+      const authContent = JSON.stringify({
+        api_key: 'sk-api-key-value',
+      });
+
+      await fs.writeFile(path.join(mockCodexDir, 'auth.json'), authContent);
+
+      const { getCodexAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getCodexAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasApiKey).toBe(true);
+    });
+  });
+
+  describe('getOpenCodeAuthIndicators', () => {
+    it('should detect provider-specific OAuth credentials', async () => {
+      const authContent = JSON.stringify({
+        anthropic: {
+          type: 'oauth',
+          access: 'oauth-access-token',
+          refresh: 'oauth-refresh-token',
+        },
+      });
+
+      await fs.writeFile(path.join(mockOpenCodeDir, 'auth.json'), authContent);
+
+      const { getOpenCodeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getOpenCodeAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(true);
+      expect(indicators.hasApiKey).toBe(false);
+    });
+
+    it('should detect GitHub Copilot refresh token as OAuth', async () => {
+      const authContent = JSON.stringify({
+        'github-copilot': {
+          type: 'oauth',
+          access: '', // Empty access token
+          refresh: 'gh-refresh-token', // But has refresh token
+        },
+      });
+
+      await fs.writeFile(path.join(mockOpenCodeDir, 'auth.json'), authContent);
+
+      const { getOpenCodeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getOpenCodeAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(true);
+    });
+
+    it('should detect provider-specific API key credentials', async () => {
+      const authContent = JSON.stringify({
+        openai: {
+          type: 'api_key',
+          key: 'sk-xxxxxxxxxxxx',
+        },
+      });
+
+      await fs.writeFile(path.join(mockOpenCodeDir, 'auth.json'), authContent);
+
+      const { getOpenCodeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getOpenCodeAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(false);
+      expect(indicators.hasApiKey).toBe(true);
+    });
+
+    it('should detect multiple providers', async () => {
+      const authContent = JSON.stringify({
+        anthropic: {
+          type: 'oauth',
+          access: 'anthropic-token',
+          refresh: 'refresh-token',
+        },
+        openai: {
+          type: 'api_key',
+          key: 'sk-xxxxxxxxxxxx',
+        },
+      });
+
+      await fs.writeFile(path.join(mockOpenCodeDir, 'auth.json'), authContent);
+
+      const { getOpenCodeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getOpenCodeAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(true);
+      expect(indicators.hasApiKey).toBe(true);
+    });
+
+    it('should handle missing OpenCode auth file', async () => {
+      // No auth file created
+      const { getOpenCodeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getOpenCodeAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(false);
+      expect(indicators.hasOAuthToken).toBe(false);
+      expect(indicators.hasApiKey).toBe(false);
+    });
+
+    it('should handle legacy top-level OAuth keys', async () => {
+      const authContent = JSON.stringify({
+        access_token: 'legacy-access-token',
+      });
+
+      await fs.writeFile(path.join(mockOpenCodeDir, 'auth.json'), authContent);
+
+      const { getOpenCodeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getOpenCodeAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(true);
+    });
+
+    it('should detect copilot provider OAuth', async () => {
+      const authContent = JSON.stringify({
+        copilot: {
+          type: 'oauth',
+          access: 'copilot-access-token',
+          refresh: 'copilot-refresh-token',
+        },
+      });
+
+      await fs.writeFile(path.join(mockOpenCodeDir, 'auth.json'), authContent);
+
+      const { getOpenCodeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getOpenCodeAuthIndicators();
+
+      expect(indicators.hasAuthFile).toBe(true);
+      expect(indicators.hasOAuthToken).toBe(true);
+    });
+  });
+
+  describe('Credential path helpers', () => {
+    it('should return correct Claude credential paths', async () => {
+      const { getClaudeCredentialPaths, getClaudeConfigDir } = await import('../src/system-paths');
+
+      const configDir = getClaudeConfigDir();
+      expect(configDir).toContain('.claude');
+
+      const credPaths = getClaudeCredentialPaths();
+      expect(credPaths.length).toBeGreaterThan(0);
+      expect(credPaths.some((p) => p.includes('.credentials.json'))).toBe(true);
+      expect(credPaths.some((p) => p.includes('credentials.json'))).toBe(true);
+    });
+
+    it('should return correct Codex auth path', async () => {
+      const { getCodexAuthPath, getCodexConfigDir } = await import('../src/system-paths');
+
+      const configDir = getCodexConfigDir();
+      expect(configDir).toContain('.codex');
+
+      const authPath = getCodexAuthPath();
+      expect(authPath).toContain('.codex');
+      expect(authPath).toContain('auth.json');
+    });
+
+    it('should return correct OpenCode auth path', async () => {
+      const { getOpenCodeAuthPath, getOpenCodeConfigDir } = await import('../src/system-paths');
+
+      const configDir = getOpenCodeConfigDir();
+      expect(configDir).toContain('opencode');
+
+      const authPath = getOpenCodeAuthPath();
+      expect(authPath).toContain('opencode');
+      expect(authPath).toContain('auth.json');
+    });
+  });
+
+  describe('Edge cases for credential detection', () => {
+    it('should handle credentials file with unexpected structure', async () => {
+      const unexpectedContent = JSON.stringify({
+        someUnexpectedKey: 'value',
+        nested: {
+          deeply: {
+            unexpected: true,
+          },
+        },
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), unexpectedContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // File with unexpected structure but no valid tokens should NOT be treated as having credentials
+      expect(indicators.hasCredentialsFile).toBe(false);
+      expect(indicators.credentials).toBeNull();
+    });
+
+    it('should handle array instead of object in credentials', async () => {
+      const arrayContent = JSON.stringify(['token1', 'token2']);
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), arrayContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Array is valid JSON but wrong structure - no valid tokens, so not treated as credentials file
+      expect(indicators.hasCredentialsFile).toBe(false);
+      expect(indicators.credentials).toBeNull();
+    });
+
+    it('should handle numeric values in credential fields', async () => {
+      const numericContent = JSON.stringify({
+        api_key: 12345,
+        oauth_token: 67890,
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), numericContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Note: Current implementation uses JavaScript truthiness which accepts numbers
+      // This documents the actual behavior - ideally would validate string type
+      expect(indicators.hasCredentialsFile).toBe(true);
+      // The implementation checks truthiness, not strict string type
+      expect(indicators.credentials?.hasOAuthToken).toBe(true);
+      expect(indicators.credentials?.hasApiKey).toBe(true);
+    });
+
+    it('should handle boolean values in credential fields', async () => {
+      const booleanContent = JSON.stringify({
+        api_key: true,
+        oauth_token: false,
+      });
+
+      await fs.writeFile(path.join(mockClaudeDir, '.credentials.json'), booleanContent);
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      // Note: Current implementation uses JavaScript truthiness
+      // api_key: true is truthy, oauth_token: false is falsy
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.credentials?.hasOAuthToken).toBe(false); // false is falsy
+      expect(indicators.credentials?.hasApiKey).toBe(true); // true is truthy
+    });
+
+    it('should handle malformed stats-cache.json gracefully', async () => {
+      await fs.writeFile(path.join(mockClaudeDir, 'stats-cache.json'), '{ invalid json }');
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasStatsCacheWithActivity).toBe(false);
+      expect(indicators.checks.statsCache.exists).toBe(true);
+      expect(indicators.checks.statsCache.error).toBeDefined();
+    });
+
+    it('should handle empty projects directory', async () => {
+      const projectsDir = path.join(mockClaudeDir, 'projects');
+      await fs.mkdir(projectsDir, { recursive: true });
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasProjectsSessions).toBe(false);
+      expect(indicators.checks.projectsDir.exists).toBe(true);
+      expect(indicators.checks.projectsDir.entryCount).toBe(0);
+    });
+  });
+
+  describe('Combined authentication scenarios', () => {
+    it('should detect CLI authenticated state with settings + sessions', async () => {
+      // Create settings file
+      await fs.writeFile(
+        path.join(mockClaudeDir, 'settings.json'),
+        JSON.stringify({ theme: 'dark' })
+      );
+
+      // Create projects directory with sessions
+      const projectsDir = path.join(mockClaudeDir, 'projects');
+      await fs.mkdir(projectsDir, { recursive: true });
+      await fs.mkdir(path.join(projectsDir, 'session-1'));
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasSettingsFile).toBe(true);
+      expect(indicators.hasProjectsSessions).toBe(true);
+    });
+
+    it('should detect recent activity indicating working auth', async () => {
+      // Create stats cache with recent activity
+      await fs.writeFile(
+        path.join(mockClaudeDir, 'stats-cache.json'),
+        JSON.stringify({
+          dailyActivity: [{ date: new Date().toISOString().split('T')[0], messagesCount: 10 }],
+        })
+      );
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasStatsCacheWithActivity).toBe(true);
+    });
+
+    it('should handle complete auth setup', async () => {
+      // Create all auth indicators
+      await fs.writeFile(
+        path.join(mockClaudeDir, '.credentials.json'),
+        JSON.stringify({
+          claudeAiOauth: {
+            accessToken: 'token',
+            refreshToken: 'refresh',
+          },
+        })
+      );
+      await fs.writeFile(
+        path.join(mockClaudeDir, 'settings.json'),
+        JSON.stringify({ theme: 'dark' })
+      );
+      await fs.writeFile(
+        path.join(mockClaudeDir, 'stats-cache.json'),
+        JSON.stringify({ dailyActivity: [{ date: '2025-01-15', messagesCount: 5 }] })
+      );
+      const projectsDir = path.join(mockClaudeDir, 'projects');
+      await fs.mkdir(projectsDir, { recursive: true });
+      await fs.mkdir(path.join(projectsDir, 'session-1'));
+
+      const { getClaudeAuthIndicators } = await import('../src/system-paths');
+      const indicators = await getClaudeAuthIndicators();
+
+      expect(indicators.hasCredentialsFile).toBe(true);
+      expect(indicators.hasSettingsFile).toBe(true);
+      expect(indicators.hasStatsCacheWithActivity).toBe(true);
+      expect(indicators.hasProjectsSessions).toBe(true);
+      expect(indicators.credentials?.hasOAuthToken).toBe(true);
+    });
+  });
+});
--- a/libs/types/src/codex-models.ts
+++ b/libs/types/src/codex-models.ts
@@ -6,6 +6,7 @@
 * IMPORTANT: All Codex models use 'codex-' prefix to distinguish from Cursor CLI models
 */
 export type CodexModelId =
+  | 'codex-gpt-5.3-codex'
  | 'codex-gpt-5.2-codex'
  | 'codex-gpt-5.1-codex-max'
  | 'codex-gpt-5.1-codex-mini'
@@ -29,31 +30,38 @@ export interface CodexModelConfig {
 * All keys use 'codex-' prefix to distinguish from Cursor CLI models
 */
 export const CODEX_MODEL_CONFIG_MAP: Record<CodexModelId, CodexModelConfig> = {
+  'codex-gpt-5.3-codex': {
+    id: 'codex-gpt-5.3-codex',
+    label: 'GPT-5.3-Codex',
+    description: 'Latest frontier agentic coding model',
+    hasThinking: true,
+    supportsVision: true,
+  },
  'codex-gpt-5.2-codex': {
    id: 'codex-gpt-5.2-codex',
    label: 'GPT-5.2-Codex',
-    description: 'Most advanced agentic coding model for complex software engineering',
+    description: 'Frontier agentic coding model',
    hasThinking: true,
    supportsVision: true,
  },
  'codex-gpt-5.1-codex-max': {
    id: 'codex-gpt-5.1-codex-max',
    label: 'GPT-5.1-Codex-Max',
-    description: 'Optimized for long-horizon, agentic coding tasks in Codex',
+    description: 'Codex-optimized flagship for deep and fast reasoning',
    hasThinking: true,
    supportsVision: true,
  },
  'codex-gpt-5.1-codex-mini': {
    id: 'codex-gpt-5.1-codex-mini',
    label: 'GPT-5.1-Codex-Mini',
-    description: 'Smaller, more cost-effective version for faster workflows',
+    description: 'Optimized for codex. Cheaper, faster, but less capable',
    hasThinking: false,
    supportsVision: true,
  },
  'codex-gpt-5.2': {
    id: 'codex-gpt-5.2',
    label: 'GPT-5.2 (Codex)',
-    description: 'Best general agentic model for tasks across industries and domains via Codex',
+    description: 'Latest frontier model with improvements across knowledge, reasoning and coding',
    hasThinking: true,
    supportsVision: true,
  },
--- a/libs/types/src/event.ts
+++ b/libs/types/src/event.ts
@@ -46,6 +46,7 @@ export type EventType =
  | 'dev-server:started'
  | 'dev-server:output'
  | 'dev-server:stopped'
+  | 'dev-server:url-detected'
  | 'test-runner:started'
  | 'test-runner:progress'
  | 'test-runner:output'
--- a/libs/types/src/index.ts
+++ b/libs/types/src/index.ts
@@ -196,6 +196,8 @@ export {
  PROJECT_SETTINGS_VERSION,
  THINKING_TOKEN_BUDGET,
  getThinkingTokenBudget,
+  isAdaptiveThinkingModel,
+  getThinkingLevelsForModel,
  // Event hook constants
  EVENT_HOOK_TRIGGER_LABELS,
  // Claude-compatible provider templates (new)
--- a/libs/types/src/model-display.ts
+++ b/libs/types/src/model-display.ts
@@ -72,10 +72,18 @@ export const CLAUDE_MODELS: ModelOption[] = [
 * Official models from https://developers.openai.com/codex/models/
 */
 export const CODEX_MODELS: (ModelOption & { hasReasoning?: boolean })[] = [
+  {
+    id: CODEX_MODEL_MAP.gpt53Codex,
+    label: 'GPT-5.3-Codex',
+    description: 'Latest frontier agentic coding model.',
+    badge: 'Premium',
+    provider: 'codex',
+    hasReasoning: true,
+  },
  {
    id: CODEX_MODEL_MAP.gpt52Codex,
    label: 'GPT-5.2-Codex',
-    description: 'Most advanced agentic coding model for complex software engineering.',
+    description: 'Frontier agentic coding model.',
    badge: 'Premium',
    provider: 'codex',
    hasReasoning: true,
@@ -83,7 +91,7 @@ export const CODEX_MODELS: (ModelOption & { hasReasoning?: boolean })[] = [
  {
    id: CODEX_MODEL_MAP.gpt51CodexMax,
    label: 'GPT-5.1-Codex-Max',
-    description: 'Optimized for long-horizon, agentic coding tasks in Codex.',
+    description: 'Codex-optimized flagship for deep and fast reasoning.',
    badge: 'Premium',
    provider: 'codex',
    hasReasoning: true,
@@ -91,7 +99,7 @@ export const CODEX_MODELS: (ModelOption & { hasReasoning?: boolean })[] = [
  {
    id: CODEX_MODEL_MAP.gpt51CodexMini,
    label: 'GPT-5.1-Codex-Mini',
-    description: 'Smaller, more cost-effective version for faster workflows.',
+    description: 'Optimized for codex. Cheaper, faster, but less capable.',
    badge: 'Speed',
    provider: 'codex',
    hasReasoning: false,
@@ -99,7 +107,7 @@ export const CODEX_MODELS: (ModelOption & { hasReasoning?: boolean })[] = [
  {
    id: CODEX_MODEL_MAP.gpt52,
    label: 'GPT-5.2',
-    description: 'Best general agentic model for tasks across industries and domains.',
+    description: 'Latest frontier model with improvements across knowledge, reasoning and coding.',
    badge: 'Balanced',
    provider: 'codex',
    hasReasoning: true,
@@ -141,6 +149,7 @@ export const THINKING_LEVELS: ThinkingLevelOption[] = [
  { id: 'medium', label: 'Medium' },
  { id: 'high', label: 'High' },
  { id: 'ultrathink', label: 'Ultrathink' },
+  { id: 'adaptive', label: 'Adaptive' },
 ];

 /**
@@ -154,6 +163,7 @@ export const THINKING_LEVEL_LABELS: Record<ThinkingLevel, string> = {
  medium: 'Med',
  high: 'High',
  ultrathink: 'Ultra',
+  adaptive: 'Adaptive',
 };

 /**
@@ -211,6 +221,7 @@ export function getModelDisplayName(model: ModelAlias | string): string {
    haiku: 'Claude Haiku',
    sonnet: 'Claude Sonnet',
    opus: 'Claude Opus',
+    [CODEX_MODEL_MAP.gpt53Codex]: 'GPT-5.3-Codex',
    [CODEX_MODEL_MAP.gpt52Codex]: 'GPT-5.2-Codex',
    [CODEX_MODEL_MAP.gpt51CodexMax]: 'GPT-5.1-Codex-Max',
    [CODEX_MODEL_MAP.gpt51CodexMini]: 'GPT-5.1-Codex-Mini',
--- a/libs/types/src/model.ts
+++ b/libs/types/src/model.ts
@@ -18,7 +18,7 @@ export type ClaudeCanonicalId = 'claude-haiku' | 'claude-sonnet' | 'claude-opus'
 export const CLAUDE_CANONICAL_MAP: Record<ClaudeCanonicalId, string> = {
  'claude-haiku': 'claude-haiku-4-5-20251001',
  'claude-sonnet': 'claude-sonnet-4-5-20250929',
-  'claude-opus': 'claude-opus-4-5-20251101',
+  'claude-opus': 'claude-opus-4-6',
 } as const;

 /**
@@ -29,7 +29,7 @@ export const CLAUDE_CANONICAL_MAP: Record<ClaudeCanonicalId, string> = {
 export const CLAUDE_MODEL_MAP: Record<string, string> = {
  haiku: 'claude-haiku-4-5-20251001',
  sonnet: 'claude-sonnet-4-5-20250929',
-  opus: 'claude-opus-4-5-20251101',
+  opus: 'claude-opus-4-6',
 } as const;

 /**
@@ -50,15 +50,17 @@ export const LEGACY_CLAUDE_ALIAS_MAP: Record<string, ClaudeCanonicalId> = {
 */
 export const CODEX_MODEL_MAP = {
  // Recommended Codex-specific models
-  /** Most advanced agentic coding model for complex software engineering (default for ChatGPT users) */
+  /** Latest frontier agentic coding model */
+  gpt53Codex: 'codex-gpt-5.3-codex',
+  /** Frontier agentic coding model */
  gpt52Codex: 'codex-gpt-5.2-codex',
-  /** Optimized for long-horizon, agentic coding tasks in Codex */
+  /** Codex-optimized flagship for deep and fast reasoning */
  gpt51CodexMax: 'codex-gpt-5.1-codex-max',
-  /** Smaller, more cost-effective version for faster workflows */
+  /** Optimized for codex. Cheaper, faster, but less capable */
  gpt51CodexMini: 'codex-gpt-5.1-codex-mini',

  // General-purpose GPT models (also available in Codex)
-  /** Best general agentic model for tasks across industries and domains */
+  /** Latest frontier model with improvements across knowledge, reasoning and coding */
  gpt52: 'codex-gpt-5.2',
  /** Great for coding and agentic tasks across domains */
  gpt51: 'codex-gpt-5.1',
@@ -71,6 +73,7 @@ export const CODEX_MODEL_IDS = Object.values(CODEX_MODEL_MAP);
 * These models can use reasoning.effort parameter
 */
 export const REASONING_CAPABLE_MODELS = new Set([
+  CODEX_MODEL_MAP.gpt53Codex,
  CODEX_MODEL_MAP.gpt52Codex,
  CODEX_MODEL_MAP.gpt51CodexMax,
  CODEX_MODEL_MAP.gpt52,
@@ -96,9 +99,9 @@ export function getAllCodexModelIds(): CodexModelId[] {
 * Uses canonical prefixed IDs for consistent routing.
 */
 export const DEFAULT_MODELS = {
-  claude: 'claude-opus-4-5-20251101',
+  claude: 'claude-opus-4-6',
  cursor: 'cursor-auto', // Cursor's recommended default (with prefix)
-  codex: CODEX_MODEL_MAP.gpt52Codex, // GPT-5.2-Codex is the most advanced agentic coding model
+  codex: CODEX_MODEL_MAP.gpt53Codex, // GPT-5.3-Codex is the latest frontier agentic coding model
 } as const;

 export type ModelAlias = keyof typeof CLAUDE_MODEL_MAP;
--- a/libs/types/src/settings.ts
+++ b/libs/types/src/settings.ts
@@ -213,7 +213,7 @@ export type PlanningMode = 'skip' | 'lite' | 'spec' | 'full';
 export type ServerLogLevel = 'error' | 'warn' | 'info' | 'debug';

 /** ThinkingLevel - Extended thinking levels for Claude models (reasoning intensity) */
-export type ThinkingLevel = 'none' | 'low' | 'medium' | 'high' | 'ultrathink';
+export type ThinkingLevel = 'none' | 'low' | 'medium' | 'high' | 'ultrathink' | 'adaptive';

 /**
 * SidebarStyle - Sidebar layout style options
@@ -237,6 +237,7 @@ export const THINKING_TOKEN_BUDGET: Record<ThinkingLevel, number | undefined> =
  medium: 10000, // Light reasoning
  high: 16000, // Complex tasks (recommended starting point)
  ultrathink: 32000, // Maximum safe (above this risks timeouts)
+  adaptive: undefined, // Adaptive thinking (Opus 4.6) - SDK handles token allocation
 };

 /**
@@ -247,6 +248,26 @@ export function getThinkingTokenBudget(level: ThinkingLevel | undefined): number
  return THINKING_TOKEN_BUDGET[level];
 }

+/**
+ * Check if a model uses adaptive thinking (Opus 4.6+)
+ * Adaptive thinking models let the SDK decide token allocation automatically.
+ */
+export function isAdaptiveThinkingModel(model: string): boolean {
+  return model.includes('opus-4-6') || model === 'claude-opus';
+}
+
+/**
+ * Get the available thinking levels for a given model.
+ * - Opus 4.6: Only 'none' and 'adaptive' (SDK handles token allocation)
+ * - Others: Full range of manual thinking levels
+ */
+export function getThinkingLevelsForModel(model: string): ThinkingLevel[] {
+  if (isAdaptiveThinkingModel(model)) {
+    return ['none', 'adaptive'];
+  }
+  return ['none', 'low', 'medium', 'high', 'ultrathink'];
+}
+
 /** ModelProvider - AI model provider for credentials and API key management */
 export type ModelProvider = 'claude' | 'cursor' | 'codex' | 'opencode' | 'gemini' | 'copilot';

--- a/package-lock.json
+++ b/package-lock.json
@@ -35,7 +35,7 @@
      "version": "0.13.0",
      "license": "SEE LICENSE IN LICENSE",
      "dependencies": {
-        "@anthropic-ai/claude-agent-sdk": "0.1.76",
+        "@anthropic-ai/claude-agent-sdk": "0.2.32",
        "@automaker/dependency-resolver": "1.0.0",
        "@automaker/git-utils": "1.0.0",
        "@automaker/model-resolver": "1.0.0",
@@ -45,7 +45,7 @@
        "@automaker/utils": "1.0.0",
        "@github/copilot-sdk": "^0.1.16",
        "@modelcontextprotocol/sdk": "1.25.2",
-        "@openai/codex-sdk": "^0.77.0",
+        "@openai/codex-sdk": "^0.98.0",
        "cookie-parser": "1.4.7",
        "cors": "2.8.5",
        "dotenv": "17.2.3",
@@ -657,9 +657,9 @@
      }
    },
    "node_modules/@anthropic-ai/claude-agent-sdk": {
-      "version": "0.1.76",
-      "resolved": "https://registry.npmjs.org/@anthropic-ai/claude-agent-sdk/-/claude-agent-sdk-0.1.76.tgz",
-      "integrity": "sha512-s7RvpXoFaLXLG7A1cJBAPD8ilwOhhc/12fb5mJXRuD561o4FmPtQ+WRfuy9akMmrFRfLsKv8Ornw3ClGAPL2fw==",
+      "version": "0.2.32",
+      "resolved": "https://registry.npmjs.org/@anthropic-ai/claude-agent-sdk/-/claude-agent-sdk-0.2.32.tgz",
+      "integrity": "sha512-8AtsSx/M9jxd0ihS08eqa7VireTEuwQy0i1+6ZJX93LECT6Svlf47dPJiAm7JB+BhVMmwTfQeS6x1akIcCfvbQ==",
      "license": "SEE LICENSE IN README.md",
      "engines": {
        "node": ">=18.0.0"
@@ -675,7 +675,7 @@
        "@img/sharp-win32-x64": "^0.33.5"
      },
      "peerDependencies": {
-        "zod": "^3.24.1 || ^4.0.0"
+        "zod": "^4.0.0"
      }
    },
    "node_modules/@automaker/dependency-resolver": {
@@ -3949,9 +3949,9 @@
      }
    },
    "node_modules/@openai/codex-sdk": {
-      "version": "0.77.0",
-      "resolved": "https://registry.npmjs.org/@openai/codex-sdk/-/codex-sdk-0.77.0.tgz",
-      "integrity": "sha512-bvJQ4dASnZ7jgfxmseViQwdRupHxs0TwHSZFeYB0gpdOAXnWwDWdGJRCMyphLSHwExRp27JNOk7EBFVmZRBanQ==",
+      "version": "0.98.0",
+      "resolved": "https://registry.npmjs.org/@openai/codex-sdk/-/codex-sdk-0.98.0.tgz",
+      "integrity": "sha512-TbPgrBpuSNMJyOXys0HNsh6UoP5VIHu1fVh2KDdACi5XyB0vuPtzBZC+qOsxHz7WXEQPFlomPLyxS6JnE5Okmg==",
      "license": "Apache-2.0",
      "engines": {
        "node": ">=18"