feat: add export command draft for session history with markdown and HTML formats

2026-01-21 08:16:21 +00:00 · 2026-01-16 17:16:02 +08:00
99 changed files with 2489 additions and 3558 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -12,7 +12,7 @@
 !.gemini/config.yaml
 !.gemini/commands/

-# Note: .qwen-clipboard/ is NOT in gitignore so Gemini can access pasted images
+# Note: .gemini-clipboard/ is NOT in gitignore so Gemini can access pasted images

 # Dependency directory
 node_modules
--- a/docs/users/configuration/auth.md
+++ b/docs/users/configuration/auth.md
@@ -5,13 +5,11 @@ Qwen Code supports two authentication methods. Pick the one that matches how you
 - **Qwen OAuth (recommended)**: sign in with your `qwen.ai` account in a browser.
 - **OpenAI-compatible API**: use an API key (OpenAI or any OpenAI-compatible provider / endpoint).

-![](https://img.alicdn.com/imgextra/i2/O1CN01IxI1bt1sNO543AVTT_!!6000000005754-0-tps-1958-822.jpg)
-
 ## Option 1: Qwen OAuth (recommended & free) 👍

-Use this if you want the simplest setup and you're using Qwen models.
+Use this if you want the simplest setup and you’re using Qwen models.

- **How it works**: on first start, Qwen Code opens a browser login page. After you finish, credentials are cached locally so you usually won't need to log in again.
+- **How it works**: on first start, Qwen Code opens a browser login page. After you finish, credentials are cached locally so you usually won’t need to log in again.
 - **Requirements**: a `qwen.ai` account + internet access (at least for the first login).
 - **Benefits**: no API key management, automatic credential refresh.
 - **Cost & quota**: free, with a quota of **60 requests/minute** and **2,000 requests/day**.
@@ -26,54 +24,15 @@ qwen

 Use this if you want to use OpenAI models or any provider that exposes an OpenAI-compatible API (e.g. OpenAI, Azure OpenAI, OpenRouter, ModelScope, Alibaba Cloud Bailian, or a self-hosted compatible endpoint).

-### Recommended: Coding Plan (subscription-based) 🚀
+### Quick start (interactive, recommended for local use)

-Use this if you want predictable costs with higher usage quotas for the qwen3-coder-plus model.
+When you choose the OpenAI-compatible option in the CLI, it will prompt you for:

-> [!IMPORTANT]
->
-> Coding Plan is only available for users in China mainland (Beijing region).
+- **API key**
+- **Base URL** (default: `https://api.openai.com/v1`)
+- **Model** (default: `gpt-4o`)

- **How it works**: subscribe to the Coding Plan with a fixed monthly fee, then configure Qwen Code to use the dedicated endpoint and your subscription API key.
- **Requirements**: an active Coding Plan subscription from [Alibaba Cloud Bailian](https://bailian.console.aliyun.com/cn-beijing/?tab=globalset#/efm/coding_plan).
- **Benefits**: higher usage quotas, predictable monthly costs, access to latest qwen3-coder-plus model.
- **Cost & quota**: varies by plan (see table below).
-
-#### Coding Plan Pricing & Quotas
-
-| Feature             | Lite Basic Plan       | Pro Advanced Plan     |
-| :------------------ | :-------------------- | :-------------------- |
-| **Price**           | ¥40/month             | ¥200/month            |
-| **5-Hour Limit**    | Up to 1,200 requests  | Up to 6,000 requests  |
-| **Weekly Limit**    | Up to 9,000 requests  | Up to 45,000 requests |
-| **Monthly Limit**   | Up to 18,000 requests | Up to 90,000 requests |
-| **Supported Model** | qwen3-coder-plus      | qwen3-coder-plus      |
-
-#### Quick Setup for Coding Plan
-
-When you select the OpenAI-compatible option in the CLI, enter these values:
-
- **API key**: `sk-sp-xxxxx`
- **Base URL**: `https://coding.dashscope.aliyuncs.com/v1`
- **Model**: `qwen3-coder-plus`
-
-> **Note**: Coding Plan API keys have the format `sk-sp-xxxxx`, which is different from standard Alibaba Cloud API keys.
-
-#### Configure via Environment Variables
-
-Set these environment variables to use Coding Plan:
-
-```bash
-export OPENAI_API_KEY="your-coding-plan-api-key"  # Format: sk-sp-xxxxx
-export OPENAI_BASE_URL="https://coding.dashscope.aliyuncs.com/v1"
-export OPENAI_MODEL="qwen3-coder-plus"
-```
-
-For more details about Coding Plan, including subscription options and troubleshooting, see the [full Coding Plan documentation](https://bailian.console.aliyun.com/cn-beijing/?tab=doc#/doc/?type=model&url=3005961).
-
-### Other OpenAI-compatible Providers
-
-If you are using other providers (OpenAI, Azure, local LLMs, etc.), use the following configuration methods.
+> **Note:** the CLI may display the key in plain text for verification. Make sure your terminal is not being recorded or shared.

 ### Configure via command-line arguments

--- a/docs/users/configuration/settings.md
+++ b/docs/users/configuration/settings.md
@@ -241,6 +241,7 @@ Per-field precedence for `generationConfig`:
 | ------------------------------------------------- | -------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ----------- |
 | `context.fileName`                                | string or array of strings | The name of the context file(s).                                                                                                                                                                                                                                                                                                                                      | `undefined` |
 | `context.importFormat`                            | string                     | The format to use when importing memory.                                                                                                                                                                                                                                                                                                                              | `undefined` |
+| `context.discoveryMaxDirs`                        | number                     | Maximum number of directories to search for memory.                                                                                                                                                                                                                                                                                                                   | `200`       |
 | `context.includeDirectories`                      | array                      | Additional directories to include in the workspace context. Specifies an array of additional absolute or relative paths to include in the workspace context. Missing directories will be skipped with a warning by default. Paths can use `~` to refer to the user's home directory. This setting can be combined with the `--include-directories` command-line flag. | `[]`        |
 | `context.loadFromIncludeDirectories`              | boolean                    | Controls the behavior of the `/memory refresh` command. If set to `true`, `QWEN.md` files should be loaded from all directories that are added. If set to `false`, `QWEN.md` should only be loaded from the current directory.                                                                                                                                        | `false`     |
 | `context.fileFiltering.respectGitIgnore`          | boolean                    | Respect .gitignore files when searching.                                                                                                                                                                                                                                                                                                                              | `true`      |
@@ -310,12 +311,6 @@ If you are experiencing performance issues with file searching (e.g., with `@` c
 >
 > **Note about advanced.tavilyApiKey:** This is a legacy configuration format. For Qwen OAuth users, DashScope provider is automatically available without any configuration. For other authentication types, configure Tavily or Google providers using the new `webSearch` configuration format.

-#### experimental
-
-| Setting               | Type    | Description                      | Default |
-| --------------------- | ------- | -------------------------------- | ------- |
-| `experimental.skills` | boolean | Enable experimental Agent Skills | `false` |
-
 #### mcpServers

 Configures connections to one or more Model-Context Protocol (MCP) servers for discovering and using custom tools. Qwen Code attempts to connect to each configured MCP server to discover available tools. If multiple MCP servers expose a tool with the same name, the tool names will be prefixed with the server alias you defined in the configuration (e.g., `serverAlias__actualToolName`) to avoid conflicts. Note that the system might strip certain schema properties from MCP tool definitions for compatibility. At least one of `command`, `url`, or `httpUrl` must be provided. If multiple are specified, the order of precedence is `httpUrl`, then `url`, then `command`.
@@ -534,13 +529,16 @@ Here's a conceptual example of what a context file at the root of a TypeScript p

 This example demonstrates how you can provide general project context, specific coding conventions, and even notes about particular files or components. The more relevant and precise your context files are, the better the AI can assist you. Project-specific context files are highly encouraged to establish conventions and context.

- **Hierarchical Loading and Precedence:** The CLI implements a hierarchical memory system by loading context files (e.g., `QWEN.md`) from several locations. Content from files lower in this list (more specific) typically overrides or supplements content from files higher up (more general). The exact concatenation order and final context can be inspected using the `/memory show` command. The typical loading order is:
+- **Hierarchical Loading and Precedence:** The CLI implements a sophisticated hierarchical memory system by loading context files (e.g., `QWEN.md`) from several locations. Content from files lower in this list (more specific) typically overrides or supplements content from files higher up (more general). The exact concatenation order and final context can be inspected using the `/memory show` command. The typical loading order is:
  1. **Global Context File:**
     - Location: `~/.qwen/<configured-context-filename>` (e.g., `~/.qwen/QWEN.md` in your user home directory).
     - Scope: Provides default instructions for all your projects.
  2. **Project Root & Ancestors Context Files:**
     - Location: The CLI searches for the configured context file in the current working directory and then in each parent directory up to either the project root (identified by a `.git` folder) or your home directory.
     - Scope: Provides context relevant to the entire project or a significant portion of it.
+  3. **Sub-directory Context Files (Contextual/Local):**
+     - Location: The CLI also scans for the configured context file in subdirectories _below_ the current working directory (respecting common ignore patterns like `node_modules`, `.git`, etc.). The breadth of this search is limited to 200 directories by default, but can be configured with the `context.discoveryMaxDirs` setting in your `settings.json` file.
+     - Scope: Allows for highly specific instructions relevant to a particular component, module, or subsection of your project.
 - **Concatenation & UI Indication:** The contents of all found context files are concatenated (with separators indicating their origin and path) and provided as part of the system prompt. The CLI footer displays the count of loaded context files, giving you a quick visual cue about the active instructional context.
 - **Importing Content:** You can modularize your context files by importing other Markdown files using the `@path/to/file.md` syntax. For more details, see the [Memory Import Processor documentation](../configuration/memory).
 - **Commands for Memory Management:**
--- a/docs/users/features/skills.md
+++ b/docs/users/features/skills.md
@@ -11,29 +11,12 @@ This guide shows you how to create, use, and manage Agent Skills in **Qwen Code*
 ## Prerequisites

 - Qwen Code (recent version)
-
-## How to enable
-
-### Via CLI flag
+- Run with the experimental flag enabled:

 ```bash
 qwen --experimental-skills
 ```

-### Via settings.json
-
-Add to your `~/.qwen/settings.json` or project's `.qwen/settings.json`:
-
-```json
-{
-  "tools": {
-    "experimental": {
-      "skills": true
-    }
-  }
-}
-```
-
 - Basic familiarity with Qwen Code ([Quickstart](../quickstart.md))

 ## What are Agent Skills?
--- a/integration-tests/acp-integration.test.ts
+++ b/integration-tests/acp-integration.test.ts
@@ -311,9 +311,9 @@ function setupAcpTest(
    }
  });

-  it('returns modes on initialize and allows setting mode and model', async () => {
+  it('returns modes on initialize and allows setting approval mode', async () => {
    const rig = new TestRig();
-    rig.setup('acp mode and model');
+    rig.setup('acp approval mode');

    const { sendRequest, cleanup, stderr } = setupAcpTest(rig);

@@ -366,14 +366,8 @@ function setupAcpTest(
      const newSession = (await sendRequest('session/new', {
        cwd: rig.testDir!,
        mcpServers: [],
-      })) as {
-        sessionId: string;
-        models: {
-          availableModels: Array<{ modelId: string }>;
-        };
-      };
+      })) as { sessionId: string };
      expect(newSession.sessionId).toBeTruthy();
-      expect(newSession.models.availableModels.length).toBeGreaterThan(0);

      // Test 4: Set approval mode to 'yolo'
      const setModeResult = (await sendRequest('session/set_mode', {
@@ -398,15 +392,6 @@ function setupAcpTest(
      })) as { modeId: string };
      expect(setModeResult3).toBeDefined();
      expect(setModeResult3.modeId).toBe('default');
-
-      // Test 7: Set model using first available model
-      const firstModel = newSession.models.availableModels[0];
-      const setModelResult = (await sendRequest('session/set_model', {
-        sessionId: newSession.sessionId,
-        modelId: firstModel.modelId,
-      })) as { modelId: string };
-      expect(setModelResult).toBeDefined();
-      expect(setModelResult.modelId).toBeTruthy();
    } catch (e) {
      if (stderr.length) {
        console.error('Agent stderr:', stderr.join(''));
--- a/package-lock.json
+++ b/package-lock.json
@@ -1,12 +1,12 @@
 {
  "name": "@qwen-code/qwen-code",
-  "version": "0.7.1",
+  "version": "0.7.0",
  "lockfileVersion": 3,
  "requires": true,
  "packages": {
    "": {
      "name": "@qwen-code/qwen-code",
-      "version": "0.7.1",
+      "version": "0.7.0",
      "workspaces": [
        "packages/*"
      ],
@@ -17310,7 +17310,7 @@
    },
    "packages/cli": {
      "name": "@qwen-code/qwen-code",
-      "version": "0.7.1",
+      "version": "0.7.0",
      "dependencies": {
        "@google/genai": "1.30.0",
        "@iarna/toml": "^2.2.5",
@@ -17947,7 +17947,7 @@
    },
    "packages/core": {
      "name": "@qwen-code/qwen-code-core",
-      "version": "0.7.1",
+      "version": "0.7.0",
      "hasInstallScript": true,
      "dependencies": {
        "@anthropic-ai/sdk": "^0.36.1",
@@ -21408,7 +21408,7 @@
    },
    "packages/test-utils": {
      "name": "@qwen-code/qwen-code-test-utils",
-      "version": "0.7.1",
+      "version": "0.7.0",
      "dev": true,
      "license": "Apache-2.0",
      "devDependencies": {
@@ -21420,7 +21420,7 @@
    },
    "packages/vscode-ide-companion": {
      "name": "qwen-code-vscode-ide-companion",
-      "version": "0.7.1",
+      "version": "0.7.0",
      "license": "LICENSE",
      "dependencies": {
        "@modelcontextprotocol/sdk": "^1.25.1",
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
 {
  "name": "@qwen-code/qwen-code",
-  "version": "0.7.1",
+  "version": "0.7.0",
  "engines": {
    "node": ">=20.0.0"
  },
@@ -13,7 +13,7 @@
    "url": "git+https://github.com/QwenLM/qwen-code.git"
  },
  "config": {
-    "sandboxImageUri": "ghcr.io/qwenlm/qwen-code:0.7.1"
+    "sandboxImageUri": "ghcr.io/qwenlm/qwen-code:0.7.0"
  },
  "scripts": {
    "start": "cross-env node scripts/start.js",
--- a/packages/cli/package.json
+++ b/packages/cli/package.json
@@ -1,6 +1,6 @@
 {
  "name": "@qwen-code/qwen-code",
-  "version": "0.7.1",
+  "version": "0.7.0",
  "description": "Qwen Code",
  "repository": {
    "type": "git",
@@ -33,7 +33,7 @@
    "dist"
  ],
  "config": {
-    "sandboxImageUri": "ghcr.io/qwenlm/qwen-code:0.7.1"
+    "sandboxImageUri": "ghcr.io/qwenlm/qwen-code:0.7.0"
  },
  "dependencies": {
    "@google/genai": "1.30.0",
--- a/packages/cli/src/acp-integration/acp.ts
+++ b/packages/cli/src/acp-integration/acp.ts
@@ -8,7 +8,6 @@

 import { z } from 'zod';
 import * as schema from './schema.js';
-import { ACP_ERROR_CODES } from './errorCodes.js';
 export * from './schema.js';

 import type { WritableStream, ReadableStream } from 'node:stream/web';
@@ -71,13 +70,6 @@ export class AgentSideConnection implements Client {
          const validatedParams = schema.setModeRequestSchema.parse(params);
          return agent.setMode(validatedParams);
        }
-        case schema.AGENT_METHODS.session_set_model: {
-          if (!agent.setModel) {
-            throw RequestError.methodNotFound();
-          }
-          const validatedParams = schema.setModelRequestSchema.parse(params);
-          return agent.setModel(validatedParams);
-        }
        default:
          throw RequestError.methodNotFound(method);
      }
@@ -350,51 +342,27 @@ export class RequestError extends Error {
  }

  static parseError(details?: string): RequestError {
-    return new RequestError(
-      ACP_ERROR_CODES.PARSE_ERROR,
-      'Parse error',
-      details,
-    );
+    return new RequestError(-32700, 'Parse error', details);
  }

  static invalidRequest(details?: string): RequestError {
-    return new RequestError(
-      ACP_ERROR_CODES.INVALID_REQUEST,
-      'Invalid request',
-      details,
-    );
+    return new RequestError(-32600, 'Invalid request', details);
  }

  static methodNotFound(details?: string): RequestError {
-    return new RequestError(
-      ACP_ERROR_CODES.METHOD_NOT_FOUND,
-      'Method not found',
-      details,
-    );
+    return new RequestError(-32601, 'Method not found', details);
  }

  static invalidParams(details?: string): RequestError {
-    return new RequestError(
-      ACP_ERROR_CODES.INVALID_PARAMS,
-      'Invalid params',
-      details,
-    );
+    return new RequestError(-32602, 'Invalid params', details);
  }

  static internalError(details?: string): RequestError {
-    return new RequestError(
-      ACP_ERROR_CODES.INTERNAL_ERROR,
-      'Internal error',
-      details,
-    );
+    return new RequestError(-32603, 'Internal error', details);
  }

  static authRequired(details?: string): RequestError {
-    return new RequestError(
-      ACP_ERROR_CODES.AUTH_REQUIRED,
-      'Authentication required',
-      details,
-    );
+    return new RequestError(-32000, 'Authentication required', details);
  }

  toResult<T>(): Result<T> {
@@ -440,5 +408,4 @@ export interface Agent {
  prompt(params: schema.PromptRequest): Promise<schema.PromptResponse>;
  cancel(params: schema.CancelNotification): Promise<void>;
  setMode?(params: schema.SetModeRequest): Promise<schema.SetModeResponse>;
-  setModel?(params: schema.SetModelRequest): Promise<schema.SetModelResponse>;
 }
--- a/packages/cli/src/acp-integration/acpAgent.ts
+++ b/packages/cli/src/acp-integration/acpAgent.ts
@@ -165,11 +165,30 @@ class GeminiAgent {
    this.setupFileSystem(config);

    const session = await this.createAndStoreSession(config);
-    const availableModels = this.buildAvailableModels(config);
+    const configuredModel = (
+      config.getModel() ||
+      this.config.getModel() ||
+      ''
+    ).trim();
+    const modelId = configuredModel || 'default';
+    const modelName = configuredModel || modelId;

    return {
      sessionId: session.getId(),
-      models: availableModels,
+      models: {
+        currentModelId: modelId,
+        availableModels: [
+          {
+            modelId,
+            name: modelName,
+            description: null,
+            _meta: {
+              contextLimit: tokenLimit(modelId),
+            },
+          },
+        ],
+        _meta: null,
+      },
    };
  }

@@ -286,29 +305,15 @@ class GeminiAgent {
  async setMode(params: acp.SetModeRequest): Promise<acp.SetModeResponse> {
    const session = this.sessions.get(params.sessionId);
    if (!session) {
-      throw acp.RequestError.invalidParams(
-        `Session not found for id: ${params.sessionId}`,
-      );
+      throw new Error(`Session not found: ${params.sessionId}`);
    }
    return session.setMode(params);
  }

-  async setModel(params: acp.SetModelRequest): Promise<acp.SetModelResponse> {
-    const session = this.sessions.get(params.sessionId);
-    if (!session) {
-      throw acp.RequestError.invalidParams(
-        `Session not found for id: ${params.sessionId}`,
-      );
-    }
-    return session.setModel(params);
-  }
-
  private async ensureAuthenticated(config: Config): Promise<void> {
    const selectedType = this.settings.merged.security?.auth?.selectedType;
    if (!selectedType) {
-      throw acp.RequestError.authRequired(
-        'Use Qwen Code CLI to authenticate first.',
-      );
+      throw acp.RequestError.authRequired('No Selected Type');
    }

    try {
@@ -377,43 +382,4 @@ class GeminiAgent {

    return session;
  }
-
-  private buildAvailableModels(
-    config: Config,
-  ): acp.NewSessionResponse['models'] {
-    const currentModelId = (
-      config.getModel() ||
-      this.config.getModel() ||
-      ''
-    ).trim();
-    const availableModels = config.getAvailableModels();
-
-    const mappedAvailableModels = availableModels.map((model) => ({
-      modelId: model.id,
-      name: model.label,
-      description: model.description ?? null,
-      _meta: {
-        contextLimit: tokenLimit(model.id),
-      },
-    }));
-
-    if (
-      currentModelId &&
-      !mappedAvailableModels.some((model) => model.modelId === currentModelId)
-    ) {
-      mappedAvailableModels.unshift({
-        modelId: currentModelId,
-        name: currentModelId,
-        description: null,
-        _meta: {
-          contextLimit: tokenLimit(currentModelId),
-        },
-      });
-    }
-
-    return {
-      currentModelId,
-      availableModels: mappedAvailableModels,
-    };
-  }
 }
--- a/packages/cli/src/acp-integration/errorCodes.ts
+++ b/packages/cli/src/acp-integration/errorCodes.ts
@@ -1,25 +0,0 @@
-/**
- * @license
- * Copyright 2025 Google LLC
- * SPDX-License-Identifier: Apache-2.0
- */
-
-export const ACP_ERROR_CODES = {
-  // Parse error: invalid JSON received by server.
-  PARSE_ERROR: -32700,
-  // Invalid request: JSON is not a valid Request object.
-  INVALID_REQUEST: -32600,
-  // Method not found: method does not exist or is unavailable.
-  METHOD_NOT_FOUND: -32601,
-  // Invalid params: invalid method parameter(s).
-  INVALID_PARAMS: -32602,
-  // Internal error: implementation-defined server error.
-  INTERNAL_ERROR: -32603,
-  // Authentication required: must authenticate before operation.
-  AUTH_REQUIRED: -32000,
-  // Resource not found: e.g. missing file.
-  RESOURCE_NOT_FOUND: -32002,
-} as const;
-
-export type AcpErrorCode =
-  (typeof ACP_ERROR_CODES)[keyof typeof ACP_ERROR_CODES];
--- a/packages/cli/src/acp-integration/schema.ts
+++ b/packages/cli/src/acp-integration/schema.ts
@@ -15,7 +15,6 @@ export const AGENT_METHODS = {
  session_prompt: 'session/prompt',
  session_list: 'session/list',
  session_set_mode: 'session/set_mode',
-  session_set_model: 'session/set_model',
 };

 export const CLIENT_METHODS = {
@@ -267,18 +266,6 @@ export const modelInfoSchema = z.object({
  name: z.string(),
 });

-export const setModelRequestSchema = z.object({
-  sessionId: z.string(),
-  modelId: z.string(),
-});
-
-export const setModelResponseSchema = z.object({
-  modelId: z.string(),
-});
-
-export type SetModelRequest = z.infer<typeof setModelRequestSchema>;
-export type SetModelResponse = z.infer<typeof setModelResponseSchema>;
-
 export const sessionModelStateSchema = z.object({
  _meta: acpMetaSchema,
  availableModels: z.array(modelInfoSchema),
@@ -605,7 +592,6 @@ export const agentResponseSchema = z.union([
  promptResponseSchema,
  listSessionsResponseSchema,
  setModeResponseSchema,
-  setModelResponseSchema,
 ]);

 export const requestPermissionRequestSchema = z.object({
@@ -638,7 +624,6 @@ export const agentRequestSchema = z.union([
  promptRequestSchema,
  listSessionsRequestSchema,
  setModeRequestSchema,
-  setModelRequestSchema,
 ]);

 export const agentNotificationSchema = sessionNotificationSchema;
--- a/packages/cli/src/acp-integration/service/filesystem.test.ts
+++ b/packages/cli/src/acp-integration/service/filesystem.test.ts
@@ -7,7 +7,6 @@
 import { describe, expect, it, vi } from 'vitest';
 import type { FileSystemService } from '@qwen-code/qwen-code-core';
 import { AcpFileSystemService } from './filesystem.js';
-import { ACP_ERROR_CODES } from '../errorCodes.js';

 const createFallback = (): FileSystemService => ({
  readTextFile: vi.fn(),
@@ -17,13 +16,11 @@ const createFallback = (): FileSystemService => ({

 describe('AcpFileSystemService', () => {
  describe('readTextFile ENOENT handling', () => {
-    it('converts RESOURCE_NOT_FOUND error to ENOENT', async () => {
-      const resourceNotFoundError = {
-        code: ACP_ERROR_CODES.RESOURCE_NOT_FOUND,
-        message: 'File not found',
-      };
+    it('parses path from ACP ENOENT message (quoted)', async () => {
      const client = {
-        readTextFile: vi.fn().mockRejectedValue(resourceNotFoundError),
+        readTextFile: vi
+          .fn()
+          .mockResolvedValue({ content: 'ERROR: ENOENT: "/remote/file.txt"' }),
      } as unknown as import('../acp.js').Client;

      const svc = new AcpFileSystemService(
@@ -33,20 +30,15 @@ describe('AcpFileSystemService', () => {
        createFallback(),
      );

-      await expect(svc.readTextFile('/some/file.txt')).rejects.toMatchObject({
+      await expect(svc.readTextFile('/local/file.txt')).rejects.toMatchObject({
        code: 'ENOENT',
-        errno: -2,
-        path: '/some/file.txt',
+        path: '/remote/file.txt',
      });
    });

-    it('re-throws other errors unchanged', async () => {
-      const otherError = {
-        code: ACP_ERROR_CODES.INTERNAL_ERROR,
-        message: 'Internal error',
-      };
+    it('falls back to requested path when none provided', async () => {
      const client = {
-        readTextFile: vi.fn().mockRejectedValue(otherError),
+        readTextFile: vi.fn().mockResolvedValue({ content: 'ERROR: ENOENT:' }),
      } as unknown as import('../acp.js').Client;

      const svc = new AcpFileSystemService(
@@ -56,34 +48,12 @@ describe('AcpFileSystemService', () => {
        createFallback(),
      );

-      await expect(svc.readTextFile('/some/file.txt')).rejects.toMatchObject({
-        code: ACP_ERROR_CODES.INTERNAL_ERROR,
-        message: 'Internal error',
+      await expect(
+        svc.readTextFile('/fallback/path.txt'),
+      ).rejects.toMatchObject({
+        code: 'ENOENT',
+        path: '/fallback/path.txt',
      });
    });
-
-    it('uses fallback when readTextFile capability is disabled', async () => {
-      const client = {
-        readTextFile: vi.fn(),
-      } as unknown as import('../acp.js').Client;
-
-      const fallback = createFallback();
-      (fallback.readTextFile as ReturnType<typeof vi.fn>).mockResolvedValue(
-        'fallback content',
-      );
-
-      const svc = new AcpFileSystemService(
-        client,
-        'session-3',
-        { readTextFile: false, writeTextFile: true },
-        fallback,
-      );
-
-      const result = await svc.readTextFile('/some/file.txt');
-
-      expect(result).toBe('fallback content');
-      expect(fallback.readTextFile).toHaveBeenCalledWith('/some/file.txt');
-      expect(client.readTextFile).not.toHaveBeenCalled();
-    });
  });
 });
--- a/packages/cli/src/acp-integration/service/filesystem.ts
+++ b/packages/cli/src/acp-integration/service/filesystem.ts
@@ -6,7 +6,6 @@

 import type { FileSystemService } from '@qwen-code/qwen-code-core';
 import type * as acp from '../acp.js';
-import { ACP_ERROR_CODES } from '../errorCodes.js';

 /**
 * ACP client-based implementation of FileSystemService
@@ -24,31 +23,25 @@ export class AcpFileSystemService implements FileSystemService {
      return this.fallback.readTextFile(filePath);
    }

-    let response: { content: string };
-    try {
-      response = await this.client.readTextFile({
-        path: filePath,
-        sessionId: this.sessionId,
-        line: null,
-        limit: null,
-      });
-    } catch (error) {
-      const errorCode =
-        typeof error === 'object' && error !== null && 'code' in error
-          ? (error as { code?: unknown }).code
-          : undefined;
+    const response = await this.client.readTextFile({
+      path: filePath,
+      sessionId: this.sessionId,
+      line: null,
+      limit: null,
+    });

-      if (errorCode === ACP_ERROR_CODES.RESOURCE_NOT_FOUND) {
-        const err = new Error(
-          `File not found: ${filePath}`,
-        ) as NodeJS.ErrnoException;
-        err.code = 'ENOENT';
-        err.errno = -2;
-        err.path = filePath;
-        throw err;
-      }
-
-      throw error;
+    if (response.content.startsWith('ERROR: ENOENT:')) {
+      // Treat ACP error strings as structured ENOENT errors without
+      // assuming a specific platform format.
+      const match = /^ERROR:\s*ENOENT:\s*(?<path>.*)$/i.exec(response.content);
+      const err = new Error(response.content) as NodeJS.ErrnoException;
+      err.code = 'ENOENT';
+      err.errno = -2;
+      const rawPath = match?.groups?.['path']?.trim();
+      err['path'] = rawPath
+        ? rawPath.replace(/^['"]|['"]$/g, '') || filePath
+        : filePath;
+      throw err;
    }

    return response.content;
--- a/packages/cli/src/acp-integration/session/Session.test.ts
+++ b/packages/cli/src/acp-integration/session/Session.test.ts
@@ -1,174 +0,0 @@
-/**
- * @license
- * Copyright 2025 Qwen
- * SPDX-License-Identifier: Apache-2.0
- */
-
-import { describe, it, expect, vi, beforeEach } from 'vitest';
-import { Session } from './Session.js';
-import type { Config, GeminiChat } from '@qwen-code/qwen-code-core';
-import { ApprovalMode } from '@qwen-code/qwen-code-core';
-import type * as acp from '../acp.js';
-import type { LoadedSettings } from '../../config/settings.js';
-import * as nonInteractiveCliCommands from '../../nonInteractiveCliCommands.js';
-
-vi.mock('../../nonInteractiveCliCommands.js', () => ({
-  getAvailableCommands: vi.fn(),
-  handleSlashCommand: vi.fn(),
-}));
-
-describe('Session', () => {
-  let mockChat: GeminiChat;
-  let mockConfig: Config;
-  let mockClient: acp.Client;
-  let mockSettings: LoadedSettings;
-  let session: Session;
-  let currentModel: string;
-  let setModelSpy: ReturnType<typeof vi.fn>;
-  let getAvailableCommandsSpy: ReturnType<typeof vi.fn>;
-
-  beforeEach(() => {
-    currentModel = 'qwen3-code-plus';
-    setModelSpy = vi.fn().mockImplementation(async (modelId: string) => {
-      currentModel = modelId;
-    });
-
-    mockChat = {
-      sendMessageStream: vi.fn(),
-      addHistory: vi.fn(),
-    } as unknown as GeminiChat;
-
-    mockConfig = {
-      setApprovalMode: vi.fn(),
-      setModel: setModelSpy,
-      getModel: vi.fn().mockImplementation(() => currentModel),
-    } as unknown as Config;
-
-    mockClient = {
-      sessionUpdate: vi.fn().mockResolvedValue(undefined),
-      requestPermission: vi.fn().mockResolvedValue({
-        outcome: { outcome: 'selected', optionId: 'proceed_once' },
-      }),
-      sendCustomNotification: vi.fn().mockResolvedValue(undefined),
-    } as unknown as acp.Client;
-
-    mockSettings = {
-      merged: {},
-    } as LoadedSettings;
-
-    getAvailableCommandsSpy = vi.mocked(nonInteractiveCliCommands)
-      .getAvailableCommands as unknown as ReturnType<typeof vi.fn>;
-    getAvailableCommandsSpy.mockResolvedValue([]);
-
-    session = new Session(
-      'test-session-id',
-      mockChat,
-      mockConfig,
-      mockClient,
-      mockSettings,
-    );
-  });
-
-  describe('setMode', () => {
-    it.each([
-      ['plan', ApprovalMode.PLAN],
-      ['default', ApprovalMode.DEFAULT],
-      ['auto-edit', ApprovalMode.AUTO_EDIT],
-      ['yolo', ApprovalMode.YOLO],
-    ] as const)('maps %s mode', async (modeId, expected) => {
-      const result = await session.setMode({
-        sessionId: 'test-session-id',
-        modeId,
-      });
-
-      expect(mockConfig.setApprovalMode).toHaveBeenCalledWith(expected);
-      expect(result).toEqual({ modeId });
-    });
-  });
-
-  describe('setModel', () => {
-    it('sets model via config and returns current model', async () => {
-      const result = await session.setModel({
-        sessionId: 'test-session-id',
-        modelId: '  qwen3-coder-plus  ',
-      });
-
-      expect(mockConfig.setModel).toHaveBeenCalledWith('qwen3-coder-plus', {
-        reason: 'user_request_acp',
-        context: 'session/set_model',
-      });
-      expect(mockConfig.getModel).toHaveBeenCalled();
-      expect(result).toEqual({ modelId: 'qwen3-coder-plus' });
-    });
-
-    it('rejects empty/whitespace model IDs', async () => {
-      await expect(
-        session.setModel({
-          sessionId: 'test-session-id',
-          modelId: '   ',
-        }),
-      ).rejects.toThrow('Invalid params');
-
-      expect(mockConfig.setModel).not.toHaveBeenCalled();
-    });
-
-    it('propagates errors from config.setModel', async () => {
-      const configError = new Error('Invalid model');
-      setModelSpy.mockRejectedValueOnce(configError);
-
-      await expect(
-        session.setModel({
-          sessionId: 'test-session-id',
-          modelId: 'invalid-model',
-        }),
-      ).rejects.toThrow('Invalid model');
-    });
-  });
-
-  describe('sendAvailableCommandsUpdate', () => {
-    it('sends available_commands_update from getAvailableCommands()', async () => {
-      getAvailableCommandsSpy.mockResolvedValueOnce([
-        {
-          name: 'init',
-          description: 'Initialize project context',
-        },
-      ]);
-
-      await session.sendAvailableCommandsUpdate();
-
-      expect(getAvailableCommandsSpy).toHaveBeenCalledWith(
-        mockConfig,
-        expect.any(AbortSignal),
-      );
-      expect(mockClient.sessionUpdate).toHaveBeenCalledWith({
-        sessionId: 'test-session-id',
-        update: {
-          sessionUpdate: 'available_commands_update',
-          availableCommands: [
-            {
-              name: 'init',
-              description: 'Initialize project context',
-              input: null,
-            },
-          ],
-        },
-      });
-    });
-
-    it('swallows errors and does not throw', async () => {
-      const consoleErrorSpy = vi
-        .spyOn(console, 'error')
-        .mockImplementation(() => undefined);
-      getAvailableCommandsSpy.mockRejectedValueOnce(
-        new Error('Command discovery failed'),
-      );
-
-      await expect(
-        session.sendAvailableCommandsUpdate(),
-      ).resolves.toBeUndefined();
-      expect(mockClient.sessionUpdate).not.toHaveBeenCalled();
-      expect(consoleErrorSpy).toHaveBeenCalled();
-      consoleErrorSpy.mockRestore();
-    });
-  });
-});
--- a/packages/cli/src/acp-integration/session/Session.ts
+++ b/packages/cli/src/acp-integration/session/Session.ts
@@ -52,8 +52,6 @@ import type {
  AvailableCommandsUpdate,
  SetModeRequest,
  SetModeResponse,
-  SetModelRequest,
-  SetModelResponse,
  ApprovalModeValue,
  CurrentModeUpdate,
 } from '../schema.js';
@@ -350,31 +348,6 @@ export class Session implements SessionContext {
    return { modeId: params.modeId };
  }

-  /**
-   * Sets the model for the current session.
-   * Validates the model ID and switches the model via Config.
-   */
-  async setModel(params: SetModelRequest): Promise<SetModelResponse> {
-    const modelId = params.modelId.trim();
-
-    if (!modelId) {
-      throw acp.RequestError.invalidParams('modelId cannot be empty');
-    }
-
-    // Attempt to set the model using config
-    await this.config.setModel(modelId, {
-      reason: 'user_request_acp',
-      context: 'session/set_model',
-    });
-
-    // Get updated model info
-    const currentModel = this.config.getModel();
-
-    return {
-      modelId: currentModel,
-    };
-  }
-
  /**
   * Sends a current_mode_update notification to the client.
   * Called after the agent switches modes (e.g., from exit_plan_mode tool).
--- a/packages/cli/src/config/config.test.ts
+++ b/packages/cli/src/config/config.test.ts
@@ -1196,6 +1196,11 @@ describe('Hierarchical Memory Loading (config.ts) - Placeholder Suite', () => {
      ],
      true,
      'tree',
+      {
+        respectGitIgnore: false,
+        respectQwenIgnore: true,
+      },
+      undefined, // maxDirs
    );
  });

--- a/packages/cli/src/config/config.ts
+++ b/packages/cli/src/config/config.ts
@@ -9,6 +9,7 @@ import {
  AuthType,
  Config,
  DEFAULT_QWEN_EMBEDDING_MODEL,
+  DEFAULT_MEMORY_FILE_FILTERING_OPTIONS,
  FileDiscoveryService,
  getCurrentGeminiMdFilename,
  loadServerHierarchicalMemory,
@@ -21,6 +22,7 @@ import {
  isToolEnabled,
  SessionService,
  type ResumedSessionData,
+  type FileFilteringOptions,
  type MCPServerConfig,
  type ToolName,
  EditTool,
@@ -332,14 +334,7 @@ export async function parseArguments(settings: Settings): Promise<CliArgs> {
        .option('experimental-skills', {
          type: 'boolean',
          description: 'Enable experimental Skills feature',
-          default: (() => {
-            const legacySkills = (
-              settings as Settings & {
-                tools?: { experimental?: { skills?: boolean } };
-              }
-            ).tools?.experimental?.skills;
-            return settings.experimental?.skills ?? legacySkills ?? false;
-          })(),
+          default: false,
        })
        .option('channel', {
          type: 'string',
@@ -648,6 +643,7 @@ export async function loadHierarchicalGeminiMemory(
  extensionContextFilePaths: string[] = [],
  folderTrust: boolean,
  memoryImportFormat: 'flat' | 'tree' = 'tree',
+  fileFilteringOptions?: FileFilteringOptions,
 ): Promise<{ memoryContent: string; fileCount: number }> {
  // FIX: Use real, canonical paths for a reliable comparison to handle symlinks.
  const realCwd = fs.realpathSync(path.resolve(currentWorkingDirectory));
@@ -673,6 +669,8 @@ export async function loadHierarchicalGeminiMemory(
    extensionContextFilePaths,
    folderTrust,
    memoryImportFormat,
+    fileFilteringOptions,
+    settings.context?.discoveryMaxDirs,
  );
 }

@@ -742,6 +740,11 @@ export async function loadCliConfig(

  const fileService = new FileDiscoveryService(cwd);

+  const fileFiltering = {
+    ...DEFAULT_MEMORY_FILE_FILTERING_OPTIONS,
+    ...settings.context?.fileFiltering,
+  };
+
  const includeDirectories = (settings.context?.includeDirectories || [])
    .map(resolvePath)
    .concat((argv.includeDirectories || []).map(resolvePath));
@@ -758,6 +761,7 @@ export async function loadCliConfig(
    extensionContextFilePaths,
    trustedFolder,
    memoryImportFormat,
+    fileFiltering,
  );

  let mcpServers = mergeMcpServers(settings, activeExtensions);
@@ -870,10 +874,11 @@ export async function loadCliConfig(
    }
  };

-  // ACP mode check: must include both --acp (current) and --experimental-acp (deprecated).
-  // Without this check, edit, write_file, run_shell_command would be excluded in ACP mode.
-  const isAcpMode = argv.acp || argv.experimentalAcp;
-  if (!interactive && !isAcpMode && inputFormat !== InputFormat.STREAM_JSON) {
+  if (
+    !interactive &&
+    !argv.experimentalAcp &&
+    inputFormat !== InputFormat.STREAM_JSON
+  ) {
    switch (approvalMode) {
      case ApprovalMode.PLAN:
      case ApprovalMode.DEFAULT:
--- a/packages/cli/src/config/keyBindings.ts
+++ b/packages/cli/src/config/keyBindings.ts
@@ -122,10 +122,9 @@ export const defaultKeyBindings: KeyBindingConfig = {

  // Auto-completion
  [Command.ACCEPT_SUGGESTION]: [{ key: 'tab' }, { key: 'return', ctrl: false }],
-  // Completion navigation uses only arrow keys
-  // Ctrl+P/N are reserved for history navigation (HISTORY_UP/DOWN)
-  [Command.COMPLETION_UP]: [{ key: 'up' }],
-  [Command.COMPLETION_DOWN]: [{ key: 'down' }],
+  // Completion navigation (arrow or Ctrl+P/N)
+  [Command.COMPLETION_UP]: [{ key: 'up' }, { key: 'p', ctrl: true }],
+  [Command.COMPLETION_DOWN]: [{ key: 'down' }, { key: 'n', ctrl: true }],

  // Text input
  // Must also exclude shift to allow shift+enter for newline
--- a/packages/cli/src/config/settings.ts
+++ b/packages/cli/src/config/settings.ts
@@ -106,6 +106,7 @@ const MIGRATION_MAP: Record<string, string> = {
  mcpServers: 'mcpServers',
  mcpServerCommand: 'mcp.serverCommand',
  memoryImportFormat: 'context.importFormat',
+  memoryDiscoveryMaxDirs: 'context.discoveryMaxDirs',
  model: 'model.name',
  preferredEditor: 'general.preferredEditor',
  sandbox: 'tools.sandbox',
@@ -921,21 +922,6 @@ export function migrateDeprecatedSettings(

      loadedSettings.setValue(scope, 'extensions', newExtensionsValue);
    }
-
-    const legacySkills = (
-      settings as Settings & {
-        tools?: { experimental?: { skills?: boolean } };
-      }
-    ).tools?.experimental?.skills;
-    if (
-      legacySkills !== undefined &&
-      settings.experimental?.skills === undefined
-    ) {
-      console.log(
-        `Migrating deprecated tools.experimental.skills setting from ${scope} settings...`,
-      );
-      loadedSettings.setValue(scope, 'experimental.skills', legacySkills);
-    }
  };

  processScope(SettingScope.User);
--- a/packages/cli/src/config/settingsSchema.ts
+++ b/packages/cli/src/config/settingsSchema.ts
@@ -434,16 +434,6 @@ const SETTINGS_SCHEMA = {
          'Show welcome back dialog when returning to a project with conversation history.',
        showInDialog: true,
      },
-      enableUserFeedback: {
-        type: 'boolean',
-        label: 'Enable User Feedback',
-        category: 'UI',
-        requiresRestart: false,
-        default: true,
-        description:
-          'Show optional feedback dialog after conversations to help improve Qwen performance.',
-        showInDialog: true,
-      },
      accessibility: {
        type: 'object',
        label: 'Accessibility',
@@ -474,15 +464,6 @@ const SETTINGS_SCHEMA = {
          },
        },
      },
-      feedbackLastShownTimestamp: {
-        type: 'number',
-        label: 'Feedback Last Shown Timestamp',
-        category: 'UI',
-        requiresRestart: false,
-        default: 0,
-        description: 'The last time the feedback dialog was shown.',
-        showInDialog: false,
-      },
    },
  },

@@ -741,6 +722,15 @@ const SETTINGS_SCHEMA = {
        description: 'The format to use when importing memory.',
        showInDialog: false,
      },
+      discoveryMaxDirs: {
+        type: 'number',
+        label: 'Memory Discovery Max Dirs',
+        category: 'Context',
+        requiresRestart: false,
+        default: 200,
+        description: 'Maximum number of directories to search for memory.',
+        showInDialog: true,
+      },
      includeDirectories: {
        type: 'array',
        label: 'Include Directories',
@@ -1217,16 +1207,6 @@ const SETTINGS_SCHEMA = {
    description: 'Setting to enable experimental features',
    showInDialog: false,
    properties: {
-      skills: {
-        type: 'boolean',
-        label: 'Skills',
-        category: 'Experimental',
-        requiresRestart: true,
-        default: false,
-        description:
-          'Enable experimental Agent Skills feature. When enabled, Qwen Code can use Skills from .qwen/skills/ and ~/.qwen/skills/.',
-        showInDialog: true,
-      },
      extensionManagement: {
        type: 'boolean',
        label: 'Extension Management',
--- a/packages/cli/src/i18n/locales/de.js
+++ b/packages/cli/src/i18n/locales/de.js
@@ -289,13 +289,6 @@ export default {
  'Show Citations': 'Quellenangaben anzeigen',
  'Custom Witty Phrases': 'Benutzerdefinierte Witzige Sprüche',
  'Enable Welcome Back': 'Willkommen-zurück aktivieren',
-  'Enable User Feedback': 'Benutzerfeedback aktivieren',
-  'How is Qwen doing this session? (optional)':
-    'Wie macht sich Qwen in dieser Sitzung? (optional)',
-  Bad: 'Schlecht',
-  Good: 'Gut',
-  'Not Sure Yet': 'Noch nicht sicher',
-  'Any other key': 'Beliebige andere Taste',
  'Disable Loading Phrases': 'Ladesprüche deaktivieren',
  'Screen Reader Mode': 'Bildschirmleser-Modus',
  'IDE Mode': 'IDE-Modus',
--- a/packages/cli/src/i18n/locales/en.js
+++ b/packages/cli/src/i18n/locales/en.js
@@ -286,13 +286,6 @@ export default {
  'Show Citations': 'Show Citations',
  'Custom Witty Phrases': 'Custom Witty Phrases',
  'Enable Welcome Back': 'Enable Welcome Back',
-  'Enable User Feedback': 'Enable User Feedback',
-  'How is Qwen doing this session? (optional)':
-    'How is Qwen doing this session? (optional)',
-  Bad: 'Bad',
-  Good: 'Good',
-  'Not Sure Yet': 'Not Sure Yet',
-  'Any other key': 'Any other key',
  'Disable Loading Phrases': 'Disable Loading Phrases',
  'Screen Reader Mode': 'Screen Reader Mode',
  'IDE Mode': 'IDE Mode',
--- a/packages/cli/src/i18n/locales/ru.js
+++ b/packages/cli/src/i18n/locales/ru.js
@@ -289,13 +289,6 @@ export default {
  'Show Citations': 'Показывать цитаты',
  'Custom Witty Phrases': 'Пользовательские остроумные фразы',
  'Enable Welcome Back': 'Включить приветствие при возврате',
-  'Enable User Feedback': 'Включить отзывы пользователей',
-  'How is Qwen doing this session? (optional)':
-    'Как дела у Qwen в этой сессии? (необязательно)',
-  Bad: 'Плохо',
-  Good: 'Хорошо',
-  'Not Sure Yet': 'Пока не уверен',
-  'Any other key': 'Любая другая клавиша',
  'Disable Loading Phrases': 'Отключить фразы при загрузке',
  'Screen Reader Mode': 'Режим программы чтения с экрана',
  'IDE Mode': 'Режим IDE',
--- a/packages/cli/src/i18n/locales/zh.js
+++ b/packages/cli/src/i18n/locales/zh.js
@@ -277,12 +277,6 @@ export default {
  'Show Citations': '显示引用',
  'Custom Witty Phrases': '自定义诙谐短语',
  'Enable Welcome Back': '启用欢迎回来',
-  'Enable User Feedback': '启用用户反馈',
-  'How is Qwen doing this session? (optional)': 'Qwen 这次表现如何？（可选）',
-  Bad: '不满意',
-  Good: '满意',
-  'Not Sure Yet': '暂不评价',
-  'Any other key': '任意其他键',
  'Disable Loading Phrases': '禁用加载短语',
  'Screen Reader Mode': '屏幕阅读器模式',
  'IDE Mode': 'IDE 模式',
@@ -879,11 +873,11 @@ export default {
  'Session Stats': '会话统计',
  'Model Usage': '模型使用情况',
  Reqs: '请求数',
-  'Input Tokens': '输入 token 数',
-  'Output Tokens': '输出 token 数',
+  'Input Tokens': '输入令牌',
+  'Output Tokens': '输出令牌',
  'Savings Highlight:': '节省亮点：',
  'of input tokens were served from the cache, reducing costs.':
-    '从缓存载入 token ，降低了成本',
+    '的输入令牌来自缓存，降低了成本',
  'Tip: For a full token breakdown, run `/stats model`.':
    '提示：要查看完整的令牌明细，请运行 `/stats model`',
  'Model Stats For Nerds': '模型统计（技术细节）',
--- a/packages/cli/src/services/BuiltinCommandLoader.ts
+++ b/packages/cli/src/services/BuiltinCommandLoader.ts
@@ -18,6 +18,7 @@ import { copyCommand } from '../ui/commands/copyCommand.js';
 import { docsCommand } from '../ui/commands/docsCommand.js';
 import { directoryCommand } from '../ui/commands/directoryCommand.js';
 import { editorCommand } from '../ui/commands/editorCommand.js';
+import { exportCommand } from '../ui/commands/exportCommand.js';
 import { extensionsCommand } from '../ui/commands/extensionsCommand.js';
 import { helpCommand } from '../ui/commands/helpCommand.js';
 import { ideCommand } from '../ui/commands/ideCommand.js';
@@ -67,6 +68,7 @@ export class BuiltinCommandLoader implements ICommandLoader {
      docsCommand,
      directoryCommand,
      editorCommand,
+      exportCommand,
      extensionsCommand,
      helpCommand,
      await ideCommand(),
--- a/packages/cli/src/ui/AppContainer.tsx
+++ b/packages/cli/src/ui/AppContainer.tsx
@@ -45,7 +45,6 @@ import process from 'node:process';
 import { useHistory } from './hooks/useHistoryManager.js';
 import { useMemoryMonitor } from './hooks/useMemoryMonitor.js';
 import { useThemeCommand } from './hooks/useThemeCommand.js';
-import { useFeedbackDialog } from './hooks/useFeedbackDialog.js';
 import { useAuthCommand } from './auth/useAuth.js';
 import { useEditorSettings } from './hooks/useEditorSettings.js';
 import { useSettingsCommand } from './hooks/useSettingsCommand.js';
@@ -576,6 +575,7 @@ export const AppContainer = (props: AppContainerProps) => {
        config.getExtensionContextFilePaths(),
        config.isTrustedFolder(),
        settings.merged.context?.importFormat || 'tree', // Use setting or default to 'tree'
+        config.getFileFilteringOptions(),
      );

      config.setUserMemory(memoryContent);
@@ -1196,19 +1196,6 @@ export const AppContainer = (props: AppContainerProps) => {
    isApprovalModeDialogOpen ||
    isResumeDialogOpen;

-  const {
-    isFeedbackDialogOpen,
-    openFeedbackDialog,
-    closeFeedbackDialog,
-    submitFeedback,
-  } = useFeedbackDialog({
-    config,
-    settings,
-    streamingState,
-    history: historyManager.history,
-    sessionStats,
-  });
-
  const pendingHistoryItems = useMemo(
    () => [...pendingSlashCommandHistoryItems, ...pendingGeminiHistoryItems],
    [pendingSlashCommandHistoryItems, pendingGeminiHistoryItems],
@@ -1305,8 +1292,6 @@ export const AppContainer = (props: AppContainerProps) => {
      // Subagent dialogs
      isSubagentCreateDialogOpen,
      isAgentsManagerDialogOpen,
-      // Feedback dialog
-      isFeedbackDialogOpen,
    }),
    [
      isThemeDialogOpen,
@@ -1397,8 +1382,6 @@ export const AppContainer = (props: AppContainerProps) => {
      // Subagent dialogs
      isSubagentCreateDialogOpen,
      isAgentsManagerDialogOpen,
-      // Feedback dialog
-      isFeedbackDialogOpen,
    ],
  );

@@ -1439,10 +1422,6 @@ export const AppContainer = (props: AppContainerProps) => {
      openResumeDialog,
      closeResumeDialog,
      handleResume,
-      // Feedback dialog
-      openFeedbackDialog,
-      closeFeedbackDialog,
-      submitFeedback,
    }),
    [
      handleThemeSelect,
@@ -1478,10 +1457,6 @@ export const AppContainer = (props: AppContainerProps) => {
      openResumeDialog,
      closeResumeDialog,
      handleResume,
-      // Feedback dialog
-      openFeedbackDialog,
-      closeFeedbackDialog,
-      submitFeedback,
    ],
  );

--- a/packages/cli/src/ui/FeedbackDialog.tsx
+++ b/packages/cli/src/ui/FeedbackDialog.tsx
@@ -1,61 +0,0 @@
-import { Box, Text } from 'ink';
-import type React from 'react';
-import { t } from '../i18n/index.js';
-import { useUIActions } from './contexts/UIActionsContext.js';
-import { useUIState } from './contexts/UIStateContext.js';
-import { useKeypress } from './hooks/useKeypress.js';
-
-const FEEDBACK_OPTIONS = {
-  GOOD: 1,
-  BAD: 2,
-  NOT_SURE: 3,
-} as const;
-
-const FEEDBACK_OPTION_KEYS = {
-  [FEEDBACK_OPTIONS.GOOD]: '1',
-  [FEEDBACK_OPTIONS.BAD]: '2',
-  [FEEDBACK_OPTIONS.NOT_SURE]: 'any',
-} as const;
-
-export const FEEDBACK_DIALOG_KEYS = ['1', '2'] as const;
-
-export const FeedbackDialog: React.FC = () => {
-  const uiState = useUIState();
-  const uiActions = useUIActions();
-
-  useKeypress(
-    (key) => {
-      if (key.name === FEEDBACK_OPTION_KEYS[FEEDBACK_OPTIONS.GOOD]) {
-        uiActions.submitFeedback(FEEDBACK_OPTIONS.GOOD);
-      } else if (key.name === FEEDBACK_OPTION_KEYS[FEEDBACK_OPTIONS.BAD]) {
-        uiActions.submitFeedback(FEEDBACK_OPTIONS.BAD);
-      } else {
-        uiActions.submitFeedback(FEEDBACK_OPTIONS.NOT_SURE);
-      }
-
-      uiActions.closeFeedbackDialog();
-    },
-    { isActive: uiState.isFeedbackDialogOpen },
-  );
-
-  return (
-    <Box flexDirection="column" marginY={1}>
-      <Box>
-        <Text color="cyan">● </Text>
-        <Text bold>{t('How is Qwen doing this session? (optional)')}</Text>
-      </Box>
-      <Box marginTop={1}>
-        <Text color="cyan">
-          {FEEDBACK_OPTION_KEYS[FEEDBACK_OPTIONS.GOOD]}:{' '}
-        </Text>
-        <Text>{t('Good')}</Text>
-        <Text> </Text>
-        <Text color="cyan">{FEEDBACK_OPTION_KEYS[FEEDBACK_OPTIONS.BAD]}: </Text>
-        <Text>{t('Bad')}</Text>
-        <Text> </Text>
-        <Text color="cyan">{t('Any other key')}: </Text>
-        <Text>{t('Not Sure Yet')}</Text>
-      </Box>
-    </Box>
-  );
-};
--- a/packages/cli/src/ui/auth/useAuth.ts
+++ b/packages/cli/src/ui/auth/useAuth.ts
@@ -4,11 +4,7 @@
 * SPDX-License-Identifier: Apache-2.0
 */

-import type {
-  Config,
-  ContentGeneratorConfig,
-  ModelProvidersConfig,
-} from '@qwen-code/qwen-code-core';
+import type { Config, ModelProvidersConfig } from '@qwen-code/qwen-code-core';
 import {
  AuthEvent,
  AuthType,
@@ -218,19 +214,11 @@ export const useAuthCommand = (

      if (authType === AuthType.USE_OPENAI) {
        if (credentials) {
-          // Pass settings.model.generationConfig to updateCredentials so it can be merged
-          // after clearing provider-sourced config. This ensures settings.json generationConfig
-          // fields (e.g., samplingParams, timeout) are preserved.
-          const settingsGenerationConfig = settings.merged.model
-            ?.generationConfig as Partial<ContentGeneratorConfig> | undefined;
-          config.updateCredentials(
-            {
-              apiKey: credentials.apiKey,
-              baseUrl: credentials.baseUrl,
-              model: credentials.model,
-            },
-            settingsGenerationConfig,
-          );
+          config.updateCredentials({
+            apiKey: credentials.apiKey,
+            baseUrl: credentials.baseUrl,
+            model: credentials.model,
+          });
          await performAuth(authType, credentials);
        }
        return;
@@ -238,13 +226,7 @@ export const useAuthCommand = (

      await performAuth(authType);
    },
-    [
-      config,
-      performAuth,
-      isProviderManagedModel,
-      onAuthError,
-      settings.merged.model?.generationConfig,
-    ],
+    [config, performAuth, isProviderManagedModel, onAuthError],
  );

  const openAuthDialog = useCallback(() => {
--- a/packages/cli/src/ui/commands/directoryCommand.test.tsx
+++ b/packages/cli/src/ui/commands/directoryCommand.test.tsx
@@ -54,7 +54,9 @@ describe('directoryCommand', () => {
      services: {
        config: mockConfig,
        settings: {
-          merged: {},
+          merged: {
+            memoryDiscoveryMaxDirs: 1000,
+          },
        },
      },
      ui: {
--- a/packages/cli/src/ui/commands/directoryCommand.tsx
+++ b/packages/cli/src/ui/commands/directoryCommand.tsx
@@ -119,6 +119,8 @@ export const directoryCommand: SlashCommand = {
                config.getFolderTrust(),
                context.services.settings.merged.context?.importFormat ||
                  'tree', // Use setting or default to 'tree'
+                config.getFileFilteringOptions(),
+                context.services.settings.merged.context?.discoveryMaxDirs,
              );
            config.setUserMemory(memoryContent);
            config.setGeminiMdFileCount(fileCount);
--- a/packages/cli/src/ui/commands/exportCommand.test.ts
+++ b/packages/cli/src/ui/commands/exportCommand.test.ts
@@ -0,0 +1,379 @@
+/**
+ * @license
+ * Copyright 2025 Qwen Team
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import * as fs from 'node:fs/promises';
+import { exportCommand } from './exportCommand.js';
+import { createMockCommandContext } from '../../test-utils/mockCommandContext.js';
+import type { ChatRecord } from '@qwen-code/qwen-code-core';
+import type { Part, Content } from '@google/genai';
+import {
+  transformToMarkdown,
+  loadHtmlTemplate,
+  prepareExportData,
+  injectDataIntoHtmlTemplate,
+  generateExportFilename,
+} from '../utils/exportUtils.js';
+
+const mockSessionServiceMocks = vi.hoisted(() => ({
+  loadLastSession: vi.fn(),
+}));
+
+vi.mock('@qwen-code/qwen-code-core', () => {
+  class SessionService {
+    constructor(_cwd: string) {}
+    async loadLastSession() {
+      return mockSessionServiceMocks.loadLastSession();
+    }
+  }
+
+  return {
+    SessionService,
+  };
+});
+
+vi.mock('../utils/exportUtils.js', () => ({
+  transformToMarkdown: vi.fn(),
+  loadHtmlTemplate: vi.fn(),
+  prepareExportData: vi.fn(),
+  injectDataIntoHtmlTemplate: vi.fn(),
+  generateExportFilename: vi.fn(),
+}));
+
+vi.mock('node:fs/promises', () => ({
+  writeFile: vi.fn(),
+}));
+
+describe('exportCommand', () => {
+  const mockSessionData = {
+    conversation: {
+      sessionId: 'test-session-id',
+      startTime: '2025-01-01T00:00:00Z',
+      messages: [
+        {
+          type: 'user',
+          message: {
+            parts: [{ text: 'Hello' }] as Part[],
+          } as Content,
+        },
+      ] as ChatRecord[],
+    },
+  };
+
+  let mockContext: ReturnType<typeof createMockCommandContext>;
+
+  beforeEach(() => {
+    vi.clearAllMocks();
+
+    mockSessionServiceMocks.loadLastSession.mockResolvedValue(mockSessionData);
+
+    mockContext = createMockCommandContext({
+      services: {
+        config: {
+          getWorkingDir: vi.fn().mockReturnValue('/test/dir'),
+          getProjectRoot: vi.fn().mockReturnValue('/test/project'),
+        },
+      },
+    });
+
+    vi.mocked(transformToMarkdown).mockReturnValue('# Test Markdown');
+    vi.mocked(loadHtmlTemplate).mockResolvedValue(
+      '<html><script id="chat-data" type="application/json">// DATA_PLACEHOLDER</script></html>',
+    );
+    vi.mocked(prepareExportData).mockReturnValue({
+      sessionId: 'test-session-id',
+      startTime: '2025-01-01T00:00:00Z',
+      messages: mockSessionData.conversation.messages,
+    });
+    vi.mocked(injectDataIntoHtmlTemplate).mockReturnValue(
+      '<html><script id="chat-data" type="application/json">{"data": "test"}</script></html>',
+    );
+    vi.mocked(generateExportFilename).mockImplementation(
+      (ext: string) => `export-2025-01-01T00-00-00-000Z.${ext}`,
+    );
+  });
+
+  afterEach(() => {
+    vi.restoreAllMocks();
+  });
+
+  describe('command structure', () => {
+    it('should have correct name and description', () => {
+      expect(exportCommand.name).toBe('export');
+      expect(exportCommand.description).toBe(
+        'Export current session message history to a file',
+      );
+    });
+
+    it('should have md and html subcommands', () => {
+      expect(exportCommand.subCommands).toHaveLength(2);
+      expect(exportCommand.subCommands?.map((c) => c.name)).toEqual([
+        'md',
+        'html',
+      ]);
+    });
+  });
+
+  describe('exportMarkdownAction', () => {
+    it('should export session to markdown file', async () => {
+      const mdCommand = exportCommand.subCommands?.find((c) => c.name === 'md');
+      if (!mdCommand?.action) {
+        throw new Error('md command not found');
+      }
+
+      const result = await mdCommand.action(mockContext, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'info',
+        content: expect.stringContaining('export-2025-01-01T00-00-00-000Z.md'),
+      });
+
+      expect(mockSessionServiceMocks.loadLastSession).toHaveBeenCalled();
+      expect(transformToMarkdown).toHaveBeenCalledWith(
+        mockSessionData.conversation.messages,
+        'test-session-id',
+        '2025-01-01T00:00:00Z',
+      );
+      expect(generateExportFilename).toHaveBeenCalledWith('md');
+      expect(fs.writeFile).toHaveBeenCalledWith(
+        expect.stringContaining('export-2025-01-01T00-00-00-000Z.md'),
+        '# Test Markdown',
+        'utf-8',
+      );
+    });
+
+    it('should return error when config is not available', async () => {
+      const contextWithoutConfig = createMockCommandContext({
+        services: {
+          config: null,
+        },
+      });
+
+      const mdCommand = exportCommand.subCommands?.find((c) => c.name === 'md');
+      if (!mdCommand?.action) {
+        throw new Error('md command not found');
+      }
+      const result = await mdCommand.action(contextWithoutConfig, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'Configuration not available.',
+      });
+    });
+
+    it('should return error when working directory cannot be determined', async () => {
+      const contextWithoutCwd = createMockCommandContext({
+        services: {
+          config: {
+            getWorkingDir: vi.fn().mockReturnValue(null),
+            getProjectRoot: vi.fn().mockReturnValue(null),
+          },
+        },
+      });
+
+      const mdCommand = exportCommand.subCommands?.find((c) => c.name === 'md');
+      if (!mdCommand || !mdCommand.action) {
+        throw new Error('md command not found');
+      }
+      const result = await mdCommand.action(contextWithoutCwd, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'Could not determine current working directory.',
+      });
+    });
+
+    it('should return error when no session is found', async () => {
+      mockSessionServiceMocks.loadLastSession.mockResolvedValue(undefined);
+
+      const mdCommand = exportCommand.subCommands?.find((c) => c.name === 'md');
+      if (!mdCommand?.action) {
+        throw new Error('md command not found');
+      }
+      const result = await mdCommand.action(mockContext, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'No active session found to export.',
+      });
+    });
+
+    it('should handle errors during export', async () => {
+      const error = new Error('File write failed');
+      vi.mocked(fs.writeFile).mockRejectedValue(error);
+
+      const mdCommand = exportCommand.subCommands?.find((c) => c.name === 'md');
+      if (!mdCommand?.action) {
+        throw new Error('md command not found');
+      }
+      const result = await mdCommand.action(mockContext, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'Failed to export session: File write failed',
+      });
+    });
+
+    it('should use project root when working dir is not available', async () => {
+      const contextWithProjectRoot = createMockCommandContext({
+        services: {
+          config: {
+            getWorkingDir: vi.fn().mockReturnValue(null),
+            getProjectRoot: vi.fn().mockReturnValue('/test/project'),
+          },
+        },
+      });
+
+      const mdCommand = exportCommand.subCommands?.find((c) => c.name === 'md');
+      if (!mdCommand?.action) {
+        throw new Error('md command not found');
+      }
+      await mdCommand.action(contextWithProjectRoot, '');
+    });
+  });
+
+  describe('exportHtmlAction', () => {
+    it('should export session to HTML file', async () => {
+      const htmlCommand = exportCommand.subCommands?.find(
+        (c) => c.name === 'html',
+      );
+      if (!htmlCommand?.action) {
+        throw new Error('html command not found');
+      }
+
+      const result = await htmlCommand.action(mockContext, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'info',
+        content: expect.stringContaining(
+          'export-2025-01-01T00-00-00-000Z.html',
+        ),
+      });
+
+      expect(mockSessionServiceMocks.loadLastSession).toHaveBeenCalled();
+      expect(loadHtmlTemplate).toHaveBeenCalled();
+      expect(prepareExportData).toHaveBeenCalledWith(
+        mockSessionData.conversation,
+      );
+      expect(injectDataIntoHtmlTemplate).toHaveBeenCalled();
+      expect(generateExportFilename).toHaveBeenCalledWith('html');
+      expect(fs.writeFile).toHaveBeenCalledWith(
+        expect.stringContaining('export-2025-01-01T00-00-00-000Z.html'),
+        expect.stringContaining('{"data": "test"}'),
+        'utf-8',
+      );
+    });
+
+    it('should return error when config is not available', async () => {
+      const contextWithoutConfig = createMockCommandContext({
+        services: {
+          config: null,
+        },
+      });
+
+      const htmlCommand = exportCommand.subCommands?.find(
+        (c) => c.name === 'html',
+      );
+      if (!htmlCommand?.action) {
+        throw new Error('html command not found');
+      }
+      const result = await htmlCommand.action(contextWithoutConfig, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'Configuration not available.',
+      });
+    });
+
+    it('should return error when working directory cannot be determined', async () => {
+      const contextWithoutCwd = createMockCommandContext({
+        services: {
+          config: {
+            getWorkingDir: vi.fn().mockReturnValue(null),
+            getProjectRoot: vi.fn().mockReturnValue(null),
+          },
+        },
+      });
+
+      const htmlCommand = exportCommand.subCommands?.find(
+        (c) => c.name === 'html',
+      );
+      if (!htmlCommand || !htmlCommand.action) {
+        throw new Error('html command not found');
+      }
+      const result = await htmlCommand.action(contextWithoutCwd, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'Could not determine current working directory.',
+      });
+    });
+
+    it('should return error when no session is found', async () => {
+      mockSessionServiceMocks.loadLastSession.mockResolvedValue(undefined);
+
+      const htmlCommand = exportCommand.subCommands?.find(
+        (c) => c.name === 'html',
+      );
+      if (!htmlCommand?.action) {
+        throw new Error('html command not found');
+      }
+      const result = await htmlCommand.action(mockContext, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'No active session found to export.',
+      });
+    });
+
+    it('should handle errors during HTML template loading', async () => {
+      const error = new Error('Failed to fetch template');
+      vi.mocked(loadHtmlTemplate).mockRejectedValue(error);
+
+      const htmlCommand = exportCommand.subCommands?.find(
+        (c) => c.name === 'html',
+      );
+      if (!htmlCommand?.action) {
+        throw new Error('html command not found');
+      }
+      const result = await htmlCommand.action(mockContext, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'Failed to export session: Failed to fetch template',
+      });
+    });
+
+    it('should handle errors during file write', async () => {
+      const error = new Error('File write failed');
+      vi.mocked(fs.writeFile).mockRejectedValue(error);
+
+      const htmlCommand = exportCommand.subCommands?.find(
+        (c) => c.name === 'html',
+      );
+      if (!htmlCommand?.action) {
+        throw new Error('html command not found');
+      }
+      const result = await htmlCommand.action(mockContext, '');
+
+      expect(result).toEqual({
+        type: 'message',
+        messageType: 'error',
+        content: 'Failed to export session: File write failed',
+      });
+    });
+  });
+});
--- a/packages/cli/src/ui/commands/exportCommand.ts
+++ b/packages/cli/src/ui/commands/exportCommand.ts
@@ -0,0 +1,177 @@
+/**
+ * @license
+ * Copyright 2025 Qwen Team
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import * as fs from 'node:fs/promises';
+import path from 'node:path';
+import {
+  type CommandContext,
+  type SlashCommand,
+  type MessageActionReturn,
+  CommandKind,
+} from './types.js';
+import { SessionService } from '@qwen-code/qwen-code-core';
+import {
+  transformToMarkdown,
+  loadHtmlTemplate,
+  prepareExportData,
+  injectDataIntoHtmlTemplate,
+  generateExportFilename,
+} from '../utils/exportUtils.js';
+
+/**
+ * Action for the 'md' subcommand - exports session to markdown.
+ */
+async function exportMarkdownAction(
+  context: CommandContext,
+): Promise<MessageActionReturn> {
+  const { services } = context;
+  const { config } = services;
+
+  if (!config) {
+    return {
+      type: 'message',
+      messageType: 'error',
+      content: 'Configuration not available.',
+    };
+  }
+
+  const cwd = config.getWorkingDir() || config.getProjectRoot();
+  if (!cwd) {
+    return {
+      type: 'message',
+      messageType: 'error',
+      content: 'Could not determine current working directory.',
+    };
+  }
+
+  try {
+    // Load the current session
+    const sessionService = new SessionService(cwd);
+    const sessionData = await sessionService.loadLastSession();
+
+    if (!sessionData) {
+      return {
+        type: 'message',
+        messageType: 'error',
+        content: 'No active session found to export.',
+      };
+    }
+
+    const { conversation } = sessionData;
+
+    const markdown = transformToMarkdown(
+      conversation.messages,
+      conversation.sessionId,
+      conversation.startTime,
+    );
+
+    const filename = generateExportFilename('md');
+    const filepath = path.join(cwd, filename);
+
+    // Write to file
+    await fs.writeFile(filepath, markdown, 'utf-8');
+
+    return {
+      type: 'message',
+      messageType: 'info',
+      content: `Session exported to markdown: ${filename}`,
+    };
+  } catch (error) {
+    return {
+      type: 'message',
+      messageType: 'error',
+      content: `Failed to export session: ${error instanceof Error ? error.message : String(error)}`,
+    };
+  }
+}
+
+/**
+ * Action for the 'html' subcommand - exports session to HTML.
+ */
+async function exportHtmlAction(
+  context: CommandContext,
+): Promise<MessageActionReturn> {
+  const { services } = context;
+  const { config } = services;
+
+  if (!config) {
+    return {
+      type: 'message',
+      messageType: 'error',
+      content: 'Configuration not available.',
+    };
+  }
+
+  const cwd = config.getWorkingDir() || config.getProjectRoot();
+  if (!cwd) {
+    return {
+      type: 'message',
+      messageType: 'error',
+      content: 'Could not determine current working directory.',
+    };
+  }
+
+  try {
+    // Load the current session
+    const sessionService = new SessionService(cwd);
+    const sessionData = await sessionService.loadLastSession();
+
+    if (!sessionData) {
+      return {
+        type: 'message',
+        messageType: 'error',
+        content: 'No active session found to export.',
+      };
+    }
+
+    const { conversation } = sessionData;
+
+    const template = await loadHtmlTemplate();
+    const exportData = prepareExportData(conversation);
+    const html = injectDataIntoHtmlTemplate(template, exportData);
+
+    const filename = generateExportFilename('html');
+    const filepath = path.join(cwd, filename);
+
+    // Write to file
+    await fs.writeFile(filepath, html, 'utf-8');
+
+    return {
+      type: 'message',
+      messageType: 'info',
+      content: `Session exported to HTML: ${filename}`,
+    };
+  } catch (error) {
+    return {
+      type: 'message',
+      messageType: 'error',
+      content: `Failed to export session: ${error instanceof Error ? error.message : String(error)}`,
+    };
+  }
+}
+
+/**
+ * Main export command with subcommands.
+ */
+export const exportCommand: SlashCommand = {
+  name: 'export',
+  description: 'Export current session message history to a file',
+  kind: CommandKind.BUILT_IN,
+  subCommands: [
+    {
+      name: 'md',
+      description: 'Export session to markdown format',
+      kind: CommandKind.BUILT_IN,
+      action: exportMarkdownAction,
+    },
+    {
+      name: 'html',
+      description: 'Export session to HTML format',
+      kind: CommandKind.BUILT_IN,
+      action: exportHtmlAction,
+    },
+  ],
+};
--- a/packages/cli/src/ui/commands/memoryCommand.test.ts
+++ b/packages/cli/src/ui/commands/memoryCommand.test.ts
@@ -299,7 +299,9 @@ describe('memoryCommand', () => {
        services: {
          config: mockConfig,
          settings: {
-            merged: {},
+            merged: {
+              memoryDiscoveryMaxDirs: 1000,
+            },
          } as LoadedSettings,
        },
      });
--- a/packages/cli/src/ui/commands/memoryCommand.ts
+++ b/packages/cli/src/ui/commands/memoryCommand.ts
@@ -315,6 +315,8 @@ export const memoryCommand: SlashCommand = {
                config.getFolderTrust(),
                context.services.settings.merged.context?.importFormat ||
                  'tree', // Use setting or default to 'tree'
+                config.getFileFilteringOptions(),
+                context.services.settings.merged.context?.discoveryMaxDirs,
              );
            config.setUserMemory(memoryContent);
            config.setGeminiMdFileCount(fileCount);
--- a/packages/cli/src/ui/components/Composer.tsx
+++ b/packages/cli/src/ui/components/Composer.tsx
@@ -26,7 +26,6 @@ import { useSettings } from '../contexts/SettingsContext.js';
 import { ApprovalMode } from '@qwen-code/qwen-code-core';
 import { StreamingState } from '../types.js';
 import { ConfigInitDisplay } from '../components/ConfigInitDisplay.js';
-import { FeedbackDialog } from '../FeedbackDialog.js';
 import { t } from '../../i18n/index.js';

 export const Composer = () => {
@@ -135,8 +134,6 @@ export const Composer = () => {
        </OverflowProvider>
      )}

-      {uiState.isFeedbackDialogOpen && <FeedbackDialog />}
-
      {uiState.isInputActive && (
        <InputPrompt
          buffer={uiState.buffer}
--- a/packages/cli/src/ui/components/InputPrompt.test.tsx
+++ b/packages/cli/src/ui/components/InputPrompt.test.tsx
@@ -33,9 +33,6 @@ vi.mock('../hooks/useCommandCompletion.js');
 vi.mock('../hooks/useInputHistory.js');
 vi.mock('../hooks/useReverseSearchCompletion.js');
 vi.mock('../utils/clipboardUtils.js');
-vi.mock('../contexts/UIStateContext.js', () => ({
-  useUIState: vi.fn(() => ({ isFeedbackDialogOpen: false })),
-}));

 const mockSlashCommands: SlashCommand[] = [
  {
@@ -281,7 +278,7 @@ describe('InputPrompt', () => {
    unmount();
  });

-  it('should call completion.navigateUp for up arrow when suggestions are showing', async () => {
+  it('should call completion.navigateUp for both up arrow and Ctrl+P when suggestions are showing', async () => {
    mockedUseCommandCompletion.mockReturnValue({
      ...mockCommandCompletion,
      showSuggestions: true,
@@ -296,22 +293,19 @@ describe('InputPrompt', () => {
    const { stdin, unmount } = renderWithProviders(<InputPrompt {...props} />);
    await wait();

-    // Test up arrow for completion navigation
+    // Test up arrow
    stdin.write('\u001B[A'); // Up arrow
    await wait();
-    expect(mockCommandCompletion.navigateUp).toHaveBeenCalledTimes(1);
-    expect(mockCommandCompletion.navigateDown).not.toHaveBeenCalled();

-    // Ctrl+P should navigate history, not completion
    stdin.write('\u0010'); // Ctrl+P
    await wait();
-    expect(mockCommandCompletion.navigateUp).toHaveBeenCalledTimes(1);
-    expect(mockInputHistory.navigateUp).toHaveBeenCalled();
+    expect(mockCommandCompletion.navigateUp).toHaveBeenCalledTimes(2);
+    expect(mockCommandCompletion.navigateDown).not.toHaveBeenCalled();

    unmount();
  });

-  it('should call completion.navigateDown for down arrow when suggestions are showing', async () => {
+  it('should call completion.navigateDown for both down arrow and Ctrl+N when suggestions are showing', async () => {
    mockedUseCommandCompletion.mockReturnValue({
      ...mockCommandCompletion,
      showSuggestions: true,
@@ -325,17 +319,14 @@ describe('InputPrompt', () => {
    const { stdin, unmount } = renderWithProviders(<InputPrompt {...props} />);
    await wait();

-    // Test down arrow for completion navigation
+    // Test down arrow
    stdin.write('\u001B[B'); // Down arrow
    await wait();
-    expect(mockCommandCompletion.navigateDown).toHaveBeenCalledTimes(1);
-    expect(mockCommandCompletion.navigateUp).not.toHaveBeenCalled();

-    // Ctrl+N should navigate history, not completion
    stdin.write('\u000E'); // Ctrl+N
    await wait();
-    expect(mockCommandCompletion.navigateDown).toHaveBeenCalledTimes(1);
-    expect(mockInputHistory.navigateDown).toHaveBeenCalled();
+    expect(mockCommandCompletion.navigateDown).toHaveBeenCalledTimes(2);
+    expect(mockCommandCompletion.navigateUp).not.toHaveBeenCalled();

    unmount();
  });
@@ -376,7 +367,7 @@ describe('InputPrompt', () => {
    it('should handle Ctrl+V when clipboard has an image', async () => {
      vi.mocked(clipboardUtils.clipboardHasImage).mockResolvedValue(true);
      vi.mocked(clipboardUtils.saveClipboardImage).mockResolvedValue(
-        '/test/.qwen-clipboard/clipboard-123.png',
+        '/test/.gemini-clipboard/clipboard-123.png',
      );

      const { stdin, unmount } = renderWithProviders(
@@ -436,7 +427,7 @@ describe('InputPrompt', () => {
    it('should insert image path at cursor position with proper spacing', async () => {
      const imagePath = path.join(
        'test',
-        '.qwen-clipboard',
+        '.gemini-clipboard',
        'clipboard-456.png',
      );
      vi.mocked(clipboardUtils.clipboardHasImage).mockResolvedValue(true);
@@ -773,8 +764,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -802,8 +791,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -831,8 +818,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -860,8 +845,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -889,8 +872,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -919,8 +900,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -948,8 +927,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -978,8 +955,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -1008,8 +983,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -1038,8 +1011,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -1068,8 +1039,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -1100,8 +1069,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -1130,8 +1097,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
@@ -1162,8 +1127,6 @@ describe('InputPrompt', () => {
        mockCommandContext,
        false,
        expect.any(Object),
-        // active parameter: completion enabled when not just navigated history
-        true,
      );

      unmount();
--- a/packages/cli/src/ui/components/InputPrompt.tsx
+++ b/packages/cli/src/ui/components/InputPrompt.tsx
@@ -36,8 +36,6 @@ import {
 import * as path from 'node:path';
 import { SCREEN_READER_USER_PREFIX } from '../textConstants.js';
 import { useShellFocusState } from '../contexts/ShellFocusContext.js';
-import { useUIState } from '../contexts/UIStateContext.js';
-import { FEEDBACK_DIALOG_KEYS } from '../FeedbackDialog.js';
 export interface InputPromptProps {
  buffer: TextBuffer;
  onSubmit: (value: string) => void;
@@ -102,7 +100,6 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
  isEmbeddedShellFocused,
 }) => {
  const isShellFocused = useShellFocusState();
-  const uiState = useUIState();
  const [justNavigatedHistory, setJustNavigatedHistory] = useState(false);
  const [escPressCount, setEscPressCount] = useState(0);
  const [showEscapePrompt, setShowEscapePrompt] = useState(false);
@@ -138,8 +135,6 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
    commandContext,
    reverseSearchActive,
    config,
-    // Suppress completion when history navigation just occurred
-    !justNavigatedHistory,
  );

  const reverseSearchCompletion = useReverseSearchCompletion(
@@ -224,9 +219,9 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
  const inputHistory = useInputHistory({
    userMessages,
    onSubmit: handleSubmitAndClear,
-    // History navigation (Ctrl+P/N) now always works since completion navigation
-    // only uses arrow keys. Only disable in shell mode.
-    isActive: !shellModeActive,
+    isActive:
+      (!completion.showSuggestions || completion.suggestions.length === 1) &&
+      !shellModeActive,
    currentQuery: buffer.text,
    onChange: customSetTextAndResetCompletionSignal,
  });
@@ -331,14 +326,6 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
        return;
      }

-      // Intercept feedback dialog option keys (1, 2) when dialog is open
-      if (
-        uiState.isFeedbackDialogOpen &&
-        (FEEDBACK_DIALOG_KEYS as readonly string[]).includes(key.name)
-      ) {
-        return;
-      }
-
      // Reset ESC count and hide prompt on any non-ESC key
      if (key.name !== 'escape') {
        if (escPressCount > 0 || showEscapePrompt) {
@@ -683,7 +670,6 @@ export const InputPrompt: React.FC<InputPromptProps> = ({
      recentPasteTime,
      commandSearchActive,
      commandSearchCompletion,
-      uiState,
    ],
  );

--- a/packages/cli/src/ui/components/ModelDialog.tsx
+++ b/packages/cli/src/ui/components/ModelDialog.tsx
@@ -275,7 +275,7 @@ export function ModelDialog({ onClose }: ModelDialogProps): React.JSX.Element {
        persistModelSelection(settings, effectiveModelId);
        persistAuthTypeSelection(settings, effectiveAuthType);

-        const baseUrl = after?.baseUrl ?? t('(default)');
+        const baseUrl = after?.baseUrl ?? '(default)';
        const maskedKey = maskApiKey(after?.apiKey);
        uiState?.historyManager.addItem(
          {
@@ -322,7 +322,7 @@ export function ModelDialog({ onClose }: ModelDialogProps): React.JSX.Element {
            <>
              <ConfigRow
                label="Base URL"
-                value={effectiveConfig?.baseUrl ?? t('(default)')}
+                value={effectiveConfig?.baseUrl ?? ''}
                badge={formatSourceBadge(sources['baseUrl'])}
              />
              <ConfigRow
--- a/packages/cli/src/ui/components/SettingsDialog.test.tsx
+++ b/packages/cli/src/ui/components/SettingsDialog.test.tsx
@@ -1331,7 +1331,9 @@ describe('SettingsDialog', () => {
          truncateToolOutputThreshold: 50000,
          truncateToolOutputLines: 1000,
        },
-        context: {},
+        context: {
+          discoveryMaxDirs: 500,
+        },
        model: {
          maxSessionTurns: 100,
          skipNextSpeakerCheck: false,
@@ -1464,6 +1466,7 @@ describe('SettingsDialog', () => {
            disableFuzzySearch: true,
          },
          loadMemoryFromIncludeDirectories: true,
+          discoveryMaxDirs: 100,
        },
      });
      const onSelect = vi.fn();
--- a/packages/cli/src/ui/contexts/UIActionsContext.tsx
+++ b/packages/cli/src/ui/contexts/UIActionsContext.tsx
@@ -66,10 +66,6 @@ export interface UIActions {
  openResumeDialog: () => void;
  closeResumeDialog: () => void;
  handleResume: (sessionId: string) => void;
-  // Feedback dialog
-  openFeedbackDialog: () => void;
-  closeFeedbackDialog: () => void;
-  submitFeedback: (rating: number) => void;
 }

 export const UIActionsContext = createContext<UIActions | null>(null);
--- a/packages/cli/src/ui/contexts/UIStateContext.tsx
+++ b/packages/cli/src/ui/contexts/UIStateContext.tsx
@@ -126,8 +126,6 @@ export interface UIState {
  // Subagent dialogs
  isSubagentCreateDialogOpen: boolean;
  isAgentsManagerDialogOpen: boolean;
-  // Feedback dialog
-  isFeedbackDialogOpen: boolean;
 }

 export const UIStateContext = createContext<UIState | null>(null);
--- a/packages/cli/src/ui/hooks/useCommandCompletion.tsx
+++ b/packages/cli/src/ui/hooks/useCommandCompletion.tsx
@@ -45,8 +45,6 @@ export function useCommandCompletion(
  commandContext: CommandContext,
  reverseSearchActive: boolean = false,
  config?: Config,
-  // When false, suppresses showing suggestions (e.g., after history navigation)
-  active: boolean = true,
 ): UseCommandCompletionReturn {
  const {
    suggestions,
@@ -154,11 +152,7 @@ export function useCommandCompletion(
  }, [suggestions, setActiveSuggestionIndex, setVisibleStartIndex]);

  useEffect(() => {
-    if (
-      completionMode === CompletionMode.IDLE ||
-      reverseSearchActive ||
-      !active
-    ) {
+    if (completionMode === CompletionMode.IDLE || reverseSearchActive) {
      resetCompletionState();
      return;
    }
@@ -169,7 +163,6 @@ export function useCommandCompletion(
    suggestions.length,
    isLoadingSuggestions,
    reverseSearchActive,
-    active,
    resetCompletionState,
    setShowSuggestions,
  ]);
--- a/packages/cli/src/ui/hooks/useFeedbackDialog.ts
+++ b/packages/cli/src/ui/hooks/useFeedbackDialog.ts
@@ -1,178 +0,0 @@
-import { useState, useCallback, useEffect } from 'react';
-import * as fs from 'node:fs';
-import {
-  type Config,
-  logUserFeedback,
-  UserFeedbackEvent,
-  type UserFeedbackRating,
-  isNodeError,
-  AuthType,
-} from '@qwen-code/qwen-code-core';
-import { StreamingState, MessageType, type HistoryItem } from '../types.js';
-import {
-  SettingScope,
-  type LoadedSettings,
-  USER_SETTINGS_PATH,
-} from '../../config/settings.js';
-import type { SessionStatsState } from '../contexts/SessionContext.js';
-import stripJsonComments from 'strip-json-comments';
-
-const FEEDBACK_SHOW_PROBABILITY = 0.25; // 25% probability of showing feedback dialog
-const MIN_TOOL_CALLS = 10; // Minimum tool calls to show feedback dialog
-const MIN_USER_MESSAGES = 5; // Minimum user messages to show feedback dialog
-
-// Fatigue mechanism constants
-const FEEDBACK_COOLDOWN_HOURS = 24; // Hours to wait before showing feedback dialog again
-
-/**
- * Check if the last message in the conversation history is an AI response
- */
-const lastMessageIsAIResponse = (history: HistoryItem[]): boolean =>
-  history.length > 0 && history[history.length - 1].type === MessageType.GEMINI;
-
-/**
- * Read feedbackLastShownTimestamp directly from the user settings file
- */
-const getFeedbackLastShownTimestampFromFile = (): number => {
-  try {
-    if (fs.existsSync(USER_SETTINGS_PATH)) {
-      const content = fs.readFileSync(USER_SETTINGS_PATH, 'utf-8');
-      const settings = JSON.parse(stripJsonComments(content));
-      return settings?.ui?.feedbackLastShownTimestamp ?? 0;
-    }
-  } catch (error) {
-    if (isNodeError(error) && error.code !== 'ENOENT') {
-      console.warn(
-        'Failed to read feedbackLastShownTimestamp from settings file:',
-        error,
-      );
-    }
-  }
-  return 0;
-};
-
-/**
- * Check if we should show the feedback dialog based on fatigue mechanism
- */
-const shouldShowFeedbackBasedOnFatigue = (): boolean => {
-  const feedbackLastShownTimestamp = getFeedbackLastShownTimestampFromFile();
-
-  const now = Date.now();
-  const timeSinceLastShown = now - feedbackLastShownTimestamp;
-  const cooldownMs = FEEDBACK_COOLDOWN_HOURS * 60 * 60 * 1000;
-
-  return timeSinceLastShown >= cooldownMs;
-};
-
-/**
- * Check if the session meets the minimum requirements for showing feedback
- * Either tool calls > 10 OR user messages > 5
- */
-const meetsMinimumSessionRequirements = (
-  sessionStats: SessionStatsState,
-): boolean => {
-  const toolCallsCount = sessionStats.metrics.tools.totalCalls;
-  const userMessagesCount = sessionStats.promptCount;
-
-  return (
-    toolCallsCount > MIN_TOOL_CALLS || userMessagesCount > MIN_USER_MESSAGES
-  );
-};
-
-export interface UseFeedbackDialogProps {
-  config: Config;
-  settings: LoadedSettings;
-  streamingState: StreamingState;
-  history: HistoryItem[];
-  sessionStats: SessionStatsState;
-}
-
-export const useFeedbackDialog = ({
-  config,
-  settings,
-  streamingState,
-  history,
-  sessionStats,
-}: UseFeedbackDialogProps) => {
-  // Feedback dialog state
-  const [isFeedbackDialogOpen, setIsFeedbackDialogOpen] = useState(false);
-
-  const openFeedbackDialog = useCallback(() => {
-    setIsFeedbackDialogOpen(true);
-
-    // Record the timestamp when feedback dialog is shown (fire and forget)
-    settings.setValue(
-      SettingScope.User,
-      'ui.feedbackLastShownTimestamp',
-      Date.now(),
-    );
-  }, [settings]);
-
-  const closeFeedbackDialog = useCallback(
-    () => setIsFeedbackDialogOpen(false),
-    [],
-  );
-
-  const submitFeedback = useCallback(
-    (rating: number) => {
-      // Create and log the feedback event
-      const feedbackEvent = new UserFeedbackEvent(
-        sessionStats.sessionId,
-        rating as UserFeedbackRating,
-        config.getModel(),
-        config.getApprovalMode(),
-      );
-
-      logUserFeedback(config, feedbackEvent);
-      closeFeedbackDialog();
-    },
-    [config, sessionStats, closeFeedbackDialog],
-  );
-
-  useEffect(() => {
-    const checkAndShowFeedback = () => {
-      if (streamingState === StreamingState.Idle && history.length > 0) {
-        // Show feedback dialog if:
-        // 1. User is authenticated via QWEN_OAUTH
-        // 2. Qwen logger is enabled (required for feedback submission)
-        // 3. User feedback is enabled in settings
-        // 4. The last message is an AI response
-        // 5. Random chance (25% probability)
-        // 6. Meets minimum requirements (tool calls > 10 OR user messages > 5)
-        // 7. Fatigue mechanism allows showing (not shown recently across sessions)
-        if (
-          config.getAuthType() !== AuthType.QWEN_OAUTH ||
-          !config.getUsageStatisticsEnabled() ||
-          settings.merged.ui?.enableUserFeedback === false ||
-          !lastMessageIsAIResponse(history) ||
-          Math.random() > FEEDBACK_SHOW_PROBABILITY ||
-          !meetsMinimumSessionRequirements(sessionStats)
-        ) {
-          return;
-        }
-
-        // Check fatigue mechanism (synchronous)
-        if (shouldShowFeedbackBasedOnFatigue()) {
-          openFeedbackDialog();
-        }
-      }
-    };
-
-    checkAndShowFeedback();
-  }, [
-    streamingState,
-    history,
-    sessionStats,
-    isFeedbackDialogOpen,
-    openFeedbackDialog,
-    settings.merged.ui?.enableUserFeedback,
-    config,
-  ]);
-
-  return {
-    isFeedbackDialogOpen,
-    openFeedbackDialog,
-    closeFeedbackDialog,
-    submitFeedback,
-  };
-};
--- a/packages/cli/src/ui/keyMatchers.test.ts
+++ b/packages/cli/src/ui/keyMatchers.test.ts
@@ -38,10 +38,10 @@ describe('keyMatchers', () => {
    [Command.NAVIGATION_DOWN]: (key: Key) => key.name === 'down',
    [Command.ACCEPT_SUGGESTION]: (key: Key) =>
      key.name === 'tab' || (key.name === 'return' && !key.ctrl),
-    // Completion navigation only uses arrow keys (not Ctrl+P/N)
-    // to allow Ctrl+P/N to always navigate history
-    [Command.COMPLETION_UP]: (key: Key) => key.name === 'up',
-    [Command.COMPLETION_DOWN]: (key: Key) => key.name === 'down',
+    [Command.COMPLETION_UP]: (key: Key) =>
+      key.name === 'up' || (key.ctrl && key.name === 'p'),
+    [Command.COMPLETION_DOWN]: (key: Key) =>
+      key.name === 'down' || (key.ctrl && key.name === 'n'),
    [Command.ESCAPE]: (key: Key) => key.name === 'escape',
    [Command.SUBMIT]: (key: Key) =>
      key.name === 'return' && !key.ctrl && !key.meta && !key.paste,
@@ -164,26 +164,14 @@ describe('keyMatchers', () => {
      negative: [createKey('return', { ctrl: true }), createKey('space')],
    },
    {
-      // Completion navigation only uses arrow keys (not Ctrl+P/N)
-      // to allow Ctrl+P/N to always navigate history
      command: Command.COMPLETION_UP,
-      positive: [createKey('up')],
-      negative: [
-        createKey('p'),
-        createKey('down'),
-        createKey('p', { ctrl: true }),
-      ],
+      positive: [createKey('up'), createKey('p', { ctrl: true })],
+      negative: [createKey('p'), createKey('down')],
    },
    {
-      // Completion navigation only uses arrow keys (not Ctrl+P/N)
-      // to allow Ctrl+P/N to always navigate history
      command: Command.COMPLETION_DOWN,
-      positive: [createKey('down')],
-      negative: [
-        createKey('n'),
-        createKey('up'),
-        createKey('n', { ctrl: true }),
-      ],
+      positive: [createKey('down'), createKey('n', { ctrl: true })],
+      negative: [createKey('n'), createKey('up')],
    },

    // Text input
--- a/packages/cli/src/ui/utils/clipboardUtils.ts
+++ b/packages/cli/src/ui/utils/clipboardUtils.ts
@@ -44,7 +44,7 @@ export async function saveClipboardImage(
    // Create a temporary directory for clipboard images within the target directory
    // This avoids security restrictions on paths outside the target directory
    const baseDir = targetDir || process.cwd();
-    const tempDir = path.join(baseDir, '.qwen-clipboard');
+    const tempDir = path.join(baseDir, '.gemini-clipboard');
    await fs.mkdir(tempDir, { recursive: true });

    // Generate a unique filename with timestamp
@@ -120,7 +120,7 @@ export async function cleanupOldClipboardImages(
 ): Promise<void> {
  try {
    const baseDir = targetDir || process.cwd();
-    const tempDir = path.join(baseDir, '.qwen-clipboard');
+    const tempDir = path.join(baseDir, '.gemini-clipboard');
    const files = await fs.readdir(tempDir);
    const oneHourAgo = Date.now() - 60 * 60 * 1000;

--- a/packages/cli/src/ui/utils/exportUtils.test.ts
+++ b/packages/cli/src/ui/utils/exportUtils.test.ts
@@ -0,0 +1,404 @@
+/**
+ * @license
+ * Copyright 2025 Qwen Team
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, it, expect, vi, beforeEach, afterEach } from 'vitest';
+import {
+  extractTextFromContent,
+  transformToMarkdown,
+  loadHtmlTemplate,
+  prepareExportData,
+  injectDataIntoHtmlTemplate,
+  generateExportFilename,
+} from './exportUtils.js';
+import type { ChatRecord } from '@qwen-code/qwen-code-core';
+import type { Part, Content } from '@google/genai';
+
+describe('exportUtils', () => {
+  describe('extractTextFromContent', () => {
+    it('should return empty string for undefined content', () => {
+      expect(extractTextFromContent(undefined)).toBe('');
+    });
+
+    it('should return empty string for content without parts', () => {
+      expect(extractTextFromContent({} as Content)).toBe('');
+    });
+
+    it('should extract text from text parts', () => {
+      const content: Content = {
+        parts: [{ text: 'Hello' }, { text: 'World' }] as Part[],
+      };
+      expect(extractTextFromContent(content)).toBe('Hello\nWorld');
+    });
+
+    it('should format function call parts', () => {
+      const content: Content = {
+        parts: [
+          {
+            functionCall: {
+              name: 'testFunction',
+              args: { param1: 'value1' },
+            },
+          },
+        ] as Part[],
+      };
+      const result = extractTextFromContent(content);
+      expect(result).toContain('[Function Call: testFunction]');
+      expect(result).toContain('"param1": "value1"');
+    });
+
+    it('should format function response parts', () => {
+      const content: Content = {
+        parts: [
+          {
+            functionResponse: {
+              name: 'testFunction',
+              response: { result: 'success' },
+            },
+          },
+        ] as Part[],
+      };
+      const result = extractTextFromContent(content);
+      expect(result).toContain('[Function Response: testFunction]');
+      expect(result).toContain('"result": "success"');
+    });
+
+    it('should handle mixed part types', () => {
+      const content: Content = {
+        parts: [
+          { text: 'Start' },
+          {
+            functionCall: {
+              name: 'call',
+              args: {},
+            },
+          },
+          { text: 'End' },
+        ] as Part[],
+      };
+      const result = extractTextFromContent(content);
+      expect(result).toContain('Start');
+      expect(result).toContain('[Function Call: call]');
+      expect(result).toContain('End');
+    });
+  });
+
+  describe('transformToMarkdown', () => {
+    const mockMessages: ChatRecord[] = [
+      {
+        uuid: 'uuid-1',
+        parentUuid: null,
+        sessionId: 'test-session-id',
+        timestamp: '2025-01-01T00:00:00Z',
+        type: 'user',
+        cwd: '/test',
+        version: '1.0.0',
+        message: {
+          parts: [{ text: 'Hello, how are you?' }] as Part[],
+        } as Content,
+      },
+      {
+        uuid: 'uuid-2',
+        parentUuid: 'uuid-1',
+        sessionId: 'test-session-id',
+        timestamp: '2025-01-01T00:00:01Z',
+        type: 'assistant',
+        cwd: '/test',
+        version: '1.0.0',
+        message: {
+          parts: [{ text: 'I am doing well, thank you!' }] as Part[],
+        } as Content,
+      },
+    ];
+
+    it('should transform messages to markdown format', () => {
+      const result = transformToMarkdown(
+        mockMessages,
+        'test-session-id',
+        '2025-01-01T00:00:00Z',
+      );
+
+      expect(result).toContain('# Chat Session Export');
+      expect(result).toContain('**Session ID**: test-session-id');
+      expect(result).toContain('**Start Time**: 2025-01-01T00:00:00Z');
+      expect(result).toContain('## User');
+      expect(result).toContain('Hello, how are you?');
+      expect(result).toContain('## Assistant');
+      expect(result).toContain('I am doing well, thank you!');
+    });
+
+    it('should include exported timestamp', () => {
+      const before = new Date().toISOString();
+      const result = transformToMarkdown(
+        mockMessages,
+        'test-session-id',
+        '2025-01-01T00:00:00Z',
+      );
+      const after = new Date().toISOString();
+
+      expect(result).toContain('**Exported**:');
+      const exportedMatch = result.match(/\*\*Exported\*\*: (.+)/);
+      expect(exportedMatch).toBeTruthy();
+      if (exportedMatch) {
+        const exportedTime = exportedMatch[1].trim();
+        expect(exportedTime >= before).toBe(true);
+        expect(exportedTime <= after).toBe(true);
+      }
+    });
+
+    it('should format tool_result messages', () => {
+      const messages: ChatRecord[] = [
+        {
+          uuid: 'uuid-3',
+          parentUuid: 'uuid-2',
+          sessionId: 'test-session-id',
+          timestamp: '2025-01-01T00:00:02Z',
+          type: 'tool_result',
+          cwd: '/test',
+          version: '1.0.0',
+          toolCallResult: {
+            resultDisplay: 'Tool output',
+          },
+          message: {
+            parts: [{ text: 'Additional info' }] as Part[],
+          } as Content,
+        },
+      ];
+
+      const result = transformToMarkdown(
+        messages,
+        'test-session-id',
+        '2025-01-01T00:00:00Z',
+      );
+
+      expect(result).toContain('## Tool Result');
+      expect(result).toContain('```');
+      expect(result).toContain('Tool output');
+      expect(result).toContain('Additional info');
+    });
+
+    it('should format tool_result with JSON resultDisplay', () => {
+      const messages: ChatRecord[] = [
+        {
+          uuid: 'uuid-4',
+          parentUuid: 'uuid-3',
+          sessionId: 'test-session-id',
+          timestamp: '2025-01-01T00:00:03Z',
+          type: 'tool_result',
+          cwd: '/test',
+          version: '1.0.0',
+          toolCallResult: {
+            resultDisplay: '{"key": "value"}',
+          },
+          message: {} as Content,
+        },
+      ];
+
+      const result = transformToMarkdown(
+        messages,
+        'test-session-id',
+        '2025-01-01T00:00:00Z',
+      );
+
+      expect(result).toContain('## Tool Result');
+      expect(result).toContain('```');
+      expect(result).toContain('"key": "value"');
+    });
+
+    it('should handle chat compression system messages', () => {
+      const messages: ChatRecord[] = [
+        {
+          uuid: 'uuid-5',
+          parentUuid: null,
+          sessionId: 'test-session-id',
+          timestamp: '2025-01-01T00:00:04Z',
+          type: 'system',
+          subtype: 'chat_compression',
+          cwd: '/test',
+          version: '1.0.0',
+          message: {} as Content,
+        },
+      ];
+
+      const result = transformToMarkdown(
+        messages,
+        'test-session-id',
+        '2025-01-01T00:00:00Z',
+      );
+
+      expect(result).toContain('_[Chat history compressed]_');
+    });
+
+    it('should skip system messages without subtype', () => {
+      const messages: ChatRecord[] = [
+        {
+          uuid: 'uuid-6',
+          parentUuid: null,
+          sessionId: 'test-session-id',
+          timestamp: '2025-01-01T00:00:05Z',
+          type: 'system',
+          cwd: '/test',
+          version: '1.0.0',
+          message: {} as Content,
+        },
+      ];
+
+      const result = transformToMarkdown(
+        messages,
+        'test-session-id',
+        '2025-01-01T00:00:00Z',
+      );
+
+      expect(result).not.toContain('## System');
+    });
+  });
+
+  describe('loadHtmlTemplate', () => {
+    beforeEach(() => {
+      vi.stubGlobal('fetch', vi.fn());
+    });
+
+    afterEach(() => {
+      vi.unstubAllGlobals();
+    });
+
+    it('should load HTML template from URL', async () => {
+      const mockTemplate = '<html><body>Test Template</body></html>';
+      const mockResponse = {
+        ok: true,
+        text: vi.fn().mockResolvedValue(mockTemplate),
+      };
+      vi.mocked(fetch).mockResolvedValue(mockResponse as unknown as Response);
+
+      const result = await loadHtmlTemplate();
+
+      expect(result).toBe(mockTemplate);
+      expect(fetch).toHaveBeenCalledWith(
+        'https://raw.githubusercontent.com/QwenLM/qwen-code/main/template_portable.html',
+      );
+    });
+
+    it('should throw error when fetch fails', async () => {
+      const mockResponse = {
+        ok: false,
+        status: 404,
+        statusText: 'Not Found',
+      };
+      vi.mocked(fetch).mockResolvedValue(mockResponse as unknown as Response);
+
+      await expect(loadHtmlTemplate()).rejects.toThrow(
+        'Failed to fetch HTML template: 404 Not Found',
+      );
+    });
+
+    it('should throw error when network request fails', async () => {
+      const networkError = new Error('Network error');
+      vi.mocked(fetch).mockRejectedValue(networkError);
+
+      await expect(loadHtmlTemplate()).rejects.toThrow(
+        'Failed to load HTML template',
+      );
+      await expect(loadHtmlTemplate()).rejects.toThrow('Network error');
+    });
+  });
+
+  describe('prepareExportData', () => {
+    it('should prepare export data from conversation', () => {
+      const conversation = {
+        sessionId: 'test-session-id',
+        startTime: '2025-01-01T00:00:00Z',
+        messages: [
+          {
+            type: 'user',
+            message: {
+              parts: [{ text: 'Hello' }] as Part[],
+            } as Content,
+          },
+        ] as ChatRecord[],
+      };
+
+      const result = prepareExportData(conversation);
+
+      expect(result).toEqual({
+        sessionId: 'test-session-id',
+        startTime: '2025-01-01T00:00:00Z',
+        messages: conversation.messages,
+      });
+    });
+  });
+
+  describe('injectDataIntoHtmlTemplate', () => {
+    it('should inject JSON data into HTML template', () => {
+      const template = `
+        <html>
+          <body>
+            <script id="chat-data" type="application/json">
+              // DATA_PLACEHOLDER: Your JSONL data will be injected here
+            </script>
+          </body>
+        </html>
+      `;
+
+      const data = {
+        sessionId: 'test-session-id',
+        startTime: '2025-01-01T00:00:00Z',
+        messages: [] as ChatRecord[],
+      };
+
+      const result = injectDataIntoHtmlTemplate(template, data);
+
+      expect(result).toContain(
+        '<script id="chat-data" type="application/json">',
+      );
+      expect(result).toContain('"sessionId": "test-session-id"');
+      expect(result).toContain('"startTime": "2025-01-01T00:00:00Z"');
+      expect(result).not.toContain('DATA_PLACEHOLDER');
+    });
+
+    it('should handle template with whitespace around placeholder', () => {
+      const template = `<script id="chat-data" type="application/json">\n// DATA_PLACEHOLDER: Your JSONL data will be injected here\n</script>`;
+
+      const data = {
+        sessionId: 'test',
+        startTime: '2025-01-01T00:00:00Z',
+        messages: [] as ChatRecord[],
+      };
+
+      const result = injectDataIntoHtmlTemplate(template, data);
+
+      expect(result).toContain('"sessionId": "test"');
+      expect(result).not.toContain('DATA_PLACEHOLDER');
+    });
+  });
+
+  describe('generateExportFilename', () => {
+    it('should generate filename with timestamp and extension', () => {
+      const filename = generateExportFilename('md');
+
+      expect(filename).toMatch(
+        /^export-\d{4}-\d{2}-\d{2}T\d{2}-\d{2}-\d{2}-\d{3}Z\.md$/,
+      );
+    });
+
+    it('should use provided extension', () => {
+      const filename1 = generateExportFilename('html');
+      const filename2 = generateExportFilename('json');
+
+      expect(filename1).toMatch(/\.html$/);
+      expect(filename2).toMatch(/\.json$/);
+    });
+
+    it('should replace colons and dots in timestamp', () => {
+      const filename = generateExportFilename('md');
+
+      expect(filename).not.toContain(':');
+      // The filename should contain a dot only for the extension
+      expect(filename.split('.').length).toBe(2);
+      // Check that timestamp part (before extension) doesn't contain dots
+      const timestampPart = filename.split('.')[0];
+      expect(timestampPart).not.toContain('.');
+    });
+  });
+});
--- a/packages/cli/src/ui/utils/exportUtils.ts
+++ b/packages/cli/src/ui/utils/exportUtils.ts
@@ -0,0 +1,167 @@
+/**
+ * @license
+ * Copyright 2025 Qwen Team
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import type { Part, Content } from '@google/genai';
+import type { ChatRecord } from '@qwen-code/qwen-code-core';
+
+const HTML_TEMPLATE_URL =
+  'https://raw.githubusercontent.com/QwenLM/qwen-code/main/template_portable.html';
+
+/**
+ * Extracts text content from a Content object's parts.
+ */
+export function extractTextFromContent(content: Content | undefined): string {
+  if (!content?.parts) return '';
+
+  const textParts: string[] = [];
+  for (const part of content.parts as Part[]) {
+    if ('text' in part) {
+      const textPart = part as { text: string };
+      textParts.push(textPart.text);
+    } else if ('functionCall' in part) {
+      const fnPart = part as { functionCall: { name: string; args: unknown } };
+      textParts.push(
+        `[Function Call: ${fnPart.functionCall.name}]\n${JSON.stringify(fnPart.functionCall.args, null, 2)}`,
+      );
+    } else if ('functionResponse' in part) {
+      const fnResPart = part as {
+        functionResponse: { name: string; response: unknown };
+      };
+      textParts.push(
+        `[Function Response: ${fnResPart.functionResponse.name}]\n${JSON.stringify(fnResPart.functionResponse.response, null, 2)}`,
+      );
+    }
+  }
+
+  return textParts.join('\n');
+}
+
+/**
+ * Transforms ChatRecord messages to markdown format.
+ */
+export function transformToMarkdown(
+  messages: ChatRecord[],
+  sessionId: string,
+  startTime: string,
+): string {
+  const lines: string[] = [];
+
+  // Add header with metadata
+  lines.push('# Chat Session Export\n');
+  lines.push(`**Session ID**: ${sessionId}\n`);
+  lines.push(`**Start Time**: ${startTime}\n`);
+  lines.push(`**Exported**: ${new Date().toISOString()}\n`);
+  lines.push('---\n');
+
+  // Process each message
+  for (const record of messages) {
+    if (record.type === 'user') {
+      lines.push('## User\n');
+      const text = extractTextFromContent(record.message);
+      lines.push(`${text}\n`);
+    } else if (record.type === 'assistant') {
+      lines.push('## Assistant\n');
+      const text = extractTextFromContent(record.message);
+      lines.push(`${text}\n`);
+    } else if (record.type === 'tool_result') {
+      lines.push('## Tool Result\n');
+      if (record.toolCallResult) {
+        const resultDisplay = record.toolCallResult.resultDisplay;
+        if (resultDisplay) {
+          lines.push('```\n');
+          lines.push(
+            typeof resultDisplay === 'string'
+              ? resultDisplay
+              : JSON.stringify(resultDisplay, null, 2),
+          );
+          lines.push('\n```\n');
+        }
+      }
+      const text = extractTextFromContent(record.message);
+      if (text) {
+        lines.push(`${text}\n`);
+      }
+    } else if (record.type === 'system') {
+      // Skip system messages or format them minimally
+      if (record.subtype === 'chat_compression') {
+        lines.push('_[Chat history compressed]_\n');
+      }
+    }
+
+    lines.push('\n');
+  }
+
+  return lines.join('');
+}
+
+/**
+ * Loads the HTML template from a remote URL via fetch.
+ * Throws an error if the fetch fails.
+ */
+export async function loadHtmlTemplate(): Promise<string> {
+  try {
+    const response = await fetch(HTML_TEMPLATE_URL);
+    if (!response.ok) {
+      throw new Error(
+        `Failed to fetch HTML template: ${response.status} ${response.statusText}`,
+      );
+    }
+    const template = await response.text();
+    return template;
+  } catch (error) {
+    throw new Error(
+      `Failed to load HTML template from ${HTML_TEMPLATE_URL}: ${
+        error instanceof Error ? error.message : String(error)
+      }`,
+    );
+  }
+}
+
+/**
+ * Prepares export data from conversation.
+ */
+export function prepareExportData(conversation: {
+  sessionId: string;
+  startTime: string;
+  messages: ChatRecord[];
+}): {
+  sessionId: string;
+  startTime: string;
+  messages: ChatRecord[];
+} {
+  return {
+    sessionId: conversation.sessionId,
+    startTime: conversation.startTime,
+    messages: conversation.messages,
+  };
+}
+
+/**
+ * Injects JSON data into the HTML template.
+ */
+export function injectDataIntoHtmlTemplate(
+  template: string,
+  data: {
+    sessionId: string;
+    startTime: string;
+    messages: ChatRecord[];
+  },
+): string {
+  const jsonData = JSON.stringify(data, null, 2);
+  const html = template.replace(
+    /<script id="chat-data" type="application\/json">\s*\/\/ DATA_PLACEHOLDER:.*?\s*<\/script>/s,
+    `<script id="chat-data" type="application/json">\n${jsonData}\n    </script>`,
+  );
+  return html;
+}
+
+/**
+ * Generates a filename with timestamp for export files.
+ */
+export function generateExportFilename(extension: string): string {
+  const timestamp = new Date().toISOString().replace(/[:.]/g, '-');
+  return `export-${timestamp}.${extension}`;
+}
--- a/packages/cli/src/utils/commentJson.test.ts
+++ b/packages/cli/src/utils/commentJson.test.ts
@@ -8,10 +8,7 @@ import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest';
 import * as fs from 'node:fs';
 import * as path from 'node:path';
 import * as os from 'node:os';
-import {
-  updateSettingsFilePreservingFormat,
-  applyUpdates,
-} from './commentJson.js';
+import { updateSettingsFilePreservingFormat } from './commentJson.js';

 describe('commentJson', () => {
  let tempDir: string;
@@ -183,18 +180,3 @@ describe('commentJson', () => {
    });
  });
 });
-
-describe('applyUpdates', () => {
-  it('should apply updates correctly', () => {
-    const original = { a: 1, b: { c: 2 } };
-    const updates = { b: { c: 3 } };
-    const result = applyUpdates(original, updates);
-    expect(result).toEqual({ a: 1, b: { c: 3 } });
-  });
-  it('should apply updates correctly when empty', () => {
-    const original = { a: 1, b: { c: 2 } };
-    const updates = { b: {} };
-    const result = applyUpdates(original, updates);
-    expect(result).toEqual({ a: 1, b: {} });
-  });
-});
--- a/packages/cli/src/utils/commentJson.ts
+++ b/packages/cli/src/utils/commentJson.ts
@@ -38,7 +38,7 @@ export function updateSettingsFilePreservingFormat(
  fs.writeFileSync(filePath, updatedContent, 'utf-8');
 }

-export function applyUpdates(
+function applyUpdates(
  current: Record<string, unknown>,
  updates: Record<string, unknown>,
 ): Record<string, unknown> {
@@ -50,7 +50,6 @@ export function applyUpdates(
      typeof value === 'object' &&
      value !== null &&
      !Array.isArray(value) &&
-      Object.keys(value).length > 0 &&
      typeof result[key] === 'object' &&
      result[key] !== null &&
      !Array.isArray(result[key])
--- a/packages/cli/src/utils/modelConfigUtils.test.ts
+++ b/packages/cli/src/utils/modelConfigUtils.test.ts
@@ -1,722 +0,0 @@
-/**
- * @license
- * Copyright 2025 Qwen Team
- * SPDX-License-Identifier: Apache-2.0
- */
-
-import { describe, it, expect, beforeEach, afterEach, vi } from 'vitest';
-import {
-  AuthType,
-  resolveModelConfig,
-  type ProviderModelConfig,
-} from '@qwen-code/qwen-code-core';
-import {
-  getAuthTypeFromEnv,
-  resolveCliGenerationConfig,
-} from './modelConfigUtils.js';
-import type { Settings } from '../config/settings.js';
-
-vi.mock('@qwen-code/qwen-code-core', async (importOriginal) => {
-  const original =
-    await importOriginal<typeof import('@qwen-code/qwen-code-core')>();
-  return {
-    ...original,
-    resolveModelConfig: vi.fn(),
-  };
-});
-
-describe('modelConfigUtils', () => {
-  describe('getAuthTypeFromEnv', () => {
-    const originalEnv = process.env;
-
-    beforeEach(() => {
-      vi.resetModules();
-      // Start with a clean env - getAuthTypeFromEnv only checks auth-related vars
-      process.env = {};
-    });
-
-    afterEach(() => {
-      process.env = originalEnv;
-    });
-
-    it('should return USE_OPENAI when all OpenAI env vars are set', () => {
-      process.env['OPENAI_API_KEY'] = 'test-key';
-      process.env['OPENAI_MODEL'] = 'gpt-4';
-      process.env['OPENAI_BASE_URL'] = 'https://api.openai.com';
-
-      expect(getAuthTypeFromEnv()).toBe(AuthType.USE_OPENAI);
-    });
-
-    it('should return undefined when OpenAI env vars are incomplete', () => {
-      process.env['OPENAI_API_KEY'] = 'test-key';
-      process.env['OPENAI_MODEL'] = 'gpt-4';
-      // Missing OPENAI_BASE_URL
-
-      expect(getAuthTypeFromEnv()).toBeUndefined();
-    });
-
-    it('should return QWEN_OAUTH when QWEN_OAUTH is set', () => {
-      process.env['QWEN_OAUTH'] = 'true';
-
-      expect(getAuthTypeFromEnv()).toBe(AuthType.QWEN_OAUTH);
-    });
-
-    it('should return USE_GEMINI when Gemini env vars are set', () => {
-      process.env['GEMINI_API_KEY'] = 'test-key';
-      process.env['GEMINI_MODEL'] = 'gemini-pro';
-
-      expect(getAuthTypeFromEnv()).toBe(AuthType.USE_GEMINI);
-    });
-
-    it('should return undefined when Gemini env vars are incomplete', () => {
-      process.env['GEMINI_API_KEY'] = 'test-key';
-      // Missing GEMINI_MODEL
-
-      expect(getAuthTypeFromEnv()).toBeUndefined();
-    });
-
-    it('should return USE_VERTEX_AI when Google env vars are set', () => {
-      process.env['GOOGLE_API_KEY'] = 'test-key';
-      process.env['GOOGLE_MODEL'] = 'vertex-model';
-
-      expect(getAuthTypeFromEnv()).toBe(AuthType.USE_VERTEX_AI);
-    });
-
-    it('should return undefined when Google env vars are incomplete', () => {
-      process.env['GOOGLE_API_KEY'] = 'test-key';
-      // Missing GOOGLE_MODEL
-
-      expect(getAuthTypeFromEnv()).toBeUndefined();
-    });
-
-    it('should return USE_ANTHROPIC when Anthropic env vars are set', () => {
-      process.env['ANTHROPIC_API_KEY'] = 'test-key';
-      process.env['ANTHROPIC_MODEL'] = 'claude-3';
-      process.env['ANTHROPIC_BASE_URL'] = 'https://api.anthropic.com';
-
-      expect(getAuthTypeFromEnv()).toBe(AuthType.USE_ANTHROPIC);
-    });
-
-    it('should return undefined when Anthropic env vars are incomplete', () => {
-      process.env['ANTHROPIC_API_KEY'] = 'test-key';
-      process.env['ANTHROPIC_MODEL'] = 'claude-3';
-      // Missing ANTHROPIC_BASE_URL
-
-      expect(getAuthTypeFromEnv()).toBeUndefined();
-    });
-
-    it('should prioritize QWEN_OAUTH over other auth types when explicitly set', () => {
-      process.env['QWEN_OAUTH'] = 'true';
-      process.env['OPENAI_API_KEY'] = 'test-key';
-      process.env['OPENAI_MODEL'] = 'gpt-4';
-      process.env['OPENAI_BASE_URL'] = 'https://api.openai.com';
-
-      // QWEN_OAUTH is checked first, so it should be returned even when other auth vars are set
-      expect(getAuthTypeFromEnv()).toBe(AuthType.QWEN_OAUTH);
-    });
-
-    it('should return undefined when no auth env vars are set', () => {
-      expect(getAuthTypeFromEnv()).toBeUndefined();
-    });
-  });
-
-  describe('resolveCliGenerationConfig', () => {
-    const originalEnv = process.env;
-    const originalConsoleWarn = console.warn;
-
-    beforeEach(() => {
-      vi.resetModules();
-      process.env = { ...originalEnv };
-      console.warn = vi.fn();
-    });
-
-    afterEach(() => {
-      process.env = originalEnv;
-      console.warn = originalConsoleWarn;
-      vi.clearAllMocks();
-    });
-
-    function makeMockSettings(overrides?: Partial<Settings>): Settings {
-      return {
-        model: { name: 'default-model' },
-        security: {
-          auth: {
-            apiKey: 'settings-api-key',
-            baseUrl: 'https://settings.example.com',
-          },
-        },
-        ...overrides,
-      } as Settings;
-    }
-
-    it('should resolve config from argv with highest precedence', () => {
-      const argv = {
-        model: 'argv-model',
-        openaiApiKey: 'argv-key',
-        openaiBaseUrl: 'https://argv.example.com',
-      };
-      const settings = makeMockSettings();
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'argv-model',
-          apiKey: 'argv-key',
-          baseUrl: 'https://argv.example.com',
-        },
-        sources: {
-          model: { kind: 'cli', detail: '--model' },
-          apiKey: { kind: 'cli', detail: '--openaiApiKey' },
-          baseUrl: { kind: 'cli', detail: '--openaiBaseUrl' },
-        },
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.model).toBe('argv-model');
-      expect(result.apiKey).toBe('argv-key');
-      expect(result.baseUrl).toBe('https://argv.example.com');
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          cli: {
-            model: 'argv-model',
-            apiKey: 'argv-key',
-            baseUrl: 'https://argv.example.com',
-          },
-        }),
-      );
-    });
-
-    it('should resolve config from settings when argv is not provided', () => {
-      const argv = {};
-      const settings = makeMockSettings({
-        model: { name: 'settings-model' },
-        security: {
-          auth: {
-            apiKey: 'settings-key',
-            baseUrl: 'https://settings.example.com',
-          },
-        },
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'settings-model',
-          apiKey: 'settings-key',
-          baseUrl: 'https://settings.example.com',
-        },
-        sources: {
-          model: { kind: 'settings', detail: 'model.name' },
-          apiKey: { kind: 'settings', detail: 'security.auth.apiKey' },
-          baseUrl: { kind: 'settings', detail: 'security.auth.baseUrl' },
-        },
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.model).toBe('settings-model');
-      expect(result.apiKey).toBe('settings-key');
-      expect(result.baseUrl).toBe('https://settings.example.com');
-    });
-
-    it('should merge generationConfig from settings', () => {
-      const argv = {};
-      const settings = makeMockSettings({
-        model: {
-          name: 'test-model',
-          generationConfig: {
-            samplingParams: {
-              temperature: 0.7,
-              max_tokens: 1000,
-            },
-            timeout: 5000,
-          } as Record<string, unknown>,
-        },
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-          samplingParams: {
-            temperature: 0.7,
-            max_tokens: 1000,
-          },
-          timeout: 5000,
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.generationConfig.samplingParams?.temperature).toBe(0.7);
-      expect(result.generationConfig.samplingParams?.max_tokens).toBe(1000);
-      expect(result.generationConfig.timeout).toBe(5000);
-    });
-
-    it('should resolve OpenAI logging from argv', () => {
-      const argv = {
-        openaiLogging: true,
-        openaiLoggingDir: '/custom/log/dir',
-      };
-      const settings = makeMockSettings();
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.generationConfig.enableOpenAILogging).toBe(true);
-      expect(result.generationConfig.openAILoggingDir).toBe('/custom/log/dir');
-    });
-
-    it('should resolve OpenAI logging from settings when argv is undefined', () => {
-      const argv = {};
-      const settings = makeMockSettings({
-        model: {
-          name: 'test-model',
-          enableOpenAILogging: true,
-          openAILoggingDir: '/settings/log/dir',
-        },
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.generationConfig.enableOpenAILogging).toBe(true);
-      expect(result.generationConfig.openAILoggingDir).toBe(
-        '/settings/log/dir',
-      );
-    });
-
-    it('should default OpenAI logging to false when not provided', () => {
-      const argv = {};
-      const settings = makeMockSettings();
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.generationConfig.enableOpenAILogging).toBe(false);
-    });
-
-    it('should find modelProvider from settings when authType and model match', () => {
-      const argv = { model: 'provider-model' };
-      const modelProvider: ProviderModelConfig = {
-        id: 'provider-model',
-        name: 'Provider Model',
-        generationConfig: {
-          samplingParams: { temperature: 0.8 },
-        },
-      };
-      const settings = makeMockSettings({
-        modelProviders: {
-          [AuthType.USE_OPENAI]: [modelProvider],
-        },
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'provider-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          modelProvider,
-        }),
-      );
-    });
-
-    it('should find modelProvider from settings.model.name when argv.model is not provided', () => {
-      const argv = {};
-      const modelProvider: ProviderModelConfig = {
-        id: 'settings-model',
-        name: 'Settings Model',
-        generationConfig: {
-          samplingParams: { temperature: 0.9 },
-        },
-      };
-      const settings = makeMockSettings({
-        model: { name: 'settings-model' },
-        modelProviders: {
-          [AuthType.USE_OPENAI]: [modelProvider],
-        },
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'settings-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          modelProvider,
-        }),
-      );
-    });
-
-    it('should not find modelProvider when authType is undefined', () => {
-      const argv = { model: 'test-model' };
-      const settings = makeMockSettings({
-        modelProviders: {
-          [AuthType.USE_OPENAI]: [{ id: 'test-model', name: 'Test Model' }],
-        },
-      });
-      const selectedAuthType = undefined;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          modelProvider: undefined,
-        }),
-      );
-    });
-
-    it('should not find modelProvider when modelProviders is not an array', () => {
-      const argv = { model: 'test-model' };
-      const settings = makeMockSettings({
-        modelProviders: {
-          [AuthType.USE_OPENAI]: null as unknown as ProviderModelConfig[],
-        },
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          modelProvider: undefined,
-        }),
-      );
-    });
-
-    it('should log warnings from resolveModelConfig', () => {
-      const argv = {};
-      const settings = makeMockSettings();
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: ['Warning 1', 'Warning 2'],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(console.warn).toHaveBeenCalledWith('Warning 1');
-      expect(console.warn).toHaveBeenCalledWith('Warning 2');
-    });
-
-    it('should use custom env when provided', () => {
-      const argv = {};
-      const settings = makeMockSettings();
-      const selectedAuthType = AuthType.USE_OPENAI;
-      const customEnv = {
-        OPENAI_API_KEY: 'custom-key',
-        OPENAI_MODEL: 'custom-model',
-      };
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'custom-model',
-          apiKey: 'custom-key',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-        env: customEnv,
-      });
-
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          env: customEnv,
-        }),
-      );
-    });
-
-    it('should use process.env when env is not provided', () => {
-      const argv = {};
-      const settings = makeMockSettings();
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          env: process.env,
-        }),
-      );
-    });
-
-    it('should return empty strings for missing model, apiKey, and baseUrl', () => {
-      const argv = {};
-      const settings = makeMockSettings();
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: '',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.model).toBe('');
-      expect(result.apiKey).toBe('');
-      expect(result.baseUrl).toBe('');
-    });
-
-    it('should merge resolved config with logging settings', () => {
-      const argv = {
-        openaiLogging: true,
-      };
-      const settings = makeMockSettings({
-        model: {
-          name: 'test-model',
-          generationConfig: {
-            timeout: 5000,
-          } as Record<string, unknown>,
-        },
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: 'test-model',
-          apiKey: 'test-key',
-          baseUrl: 'https://test.com',
-          samplingParams: { temperature: 0.5 },
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.generationConfig).toEqual({
-        model: 'test-model',
-        apiKey: 'test-key',
-        baseUrl: 'https://test.com',
-        samplingParams: { temperature: 0.5 },
-        enableOpenAILogging: true,
-        openAILoggingDir: undefined,
-      });
-    });
-
-    it('should handle settings without model property', () => {
-      const argv = {};
-      const settings = makeMockSettings({
-        model: undefined as unknown as Settings['model'],
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: '',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      const result = resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(result.model).toBe('');
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          settings: expect.objectContaining({
-            model: undefined,
-          }),
-        }),
-      );
-    });
-
-    it('should handle settings without security.auth property', () => {
-      const argv = {};
-      const settings = makeMockSettings({
-        security: undefined,
-      });
-      const selectedAuthType = AuthType.USE_OPENAI;
-
-      vi.mocked(resolveModelConfig).mockReturnValue({
-        config: {
-          model: '',
-          apiKey: '',
-          baseUrl: '',
-        },
-        sources: {},
-        warnings: [],
-      });
-
-      resolveCliGenerationConfig({
-        argv,
-        settings,
-        selectedAuthType,
-      });
-
-      expect(vi.mocked(resolveModelConfig)).toHaveBeenCalledWith(
-        expect.objectContaining({
-          settings: expect.objectContaining({
-            apiKey: undefined,
-            baseUrl: undefined,
-          }),
-        }),
-      );
-    });
-  });
-});
--- a/packages/cli/src/utils/modelConfigUtils.ts
+++ b/packages/cli/src/utils/modelConfigUtils.ts
@@ -44,31 +44,20 @@ export interface ResolvedCliGenerationConfig {
 }

 export function getAuthTypeFromEnv(): AuthType | undefined {
+  if (process.env['OPENAI_API_KEY']) {
+    return AuthType.USE_OPENAI;
+  }
  if (process.env['QWEN_OAUTH']) {
    return AuthType.QWEN_OAUTH;
  }

-  if (
-    process.env['OPENAI_API_KEY'] &&
-    process.env['OPENAI_MODEL'] &&
-    process.env['OPENAI_BASE_URL']
-  ) {
-    return AuthType.USE_OPENAI;
-  }
-
-  if (process.env['GEMINI_API_KEY'] && process.env['GEMINI_MODEL']) {
+  if (process.env['GEMINI_API_KEY']) {
    return AuthType.USE_GEMINI;
  }
-
-  if (process.env['GOOGLE_API_KEY'] && process.env['GOOGLE_MODEL']) {
+  if (process.env['GOOGLE_API_KEY']) {
    return AuthType.USE_VERTEX_AI;
  }
-
-  if (
-    process.env['ANTHROPIC_API_KEY'] &&
-    process.env['ANTHROPIC_MODEL'] &&
-    process.env['ANTHROPIC_BASE_URL']
-  ) {
+  if (process.env['ANTHROPIC_API_KEY']) {
    return AuthType.USE_ANTHROPIC;
  }

--- a/packages/core/package.json
+++ b/packages/core/package.json
@@ -1,6 +1,6 @@
 {
  "name": "@qwen-code/qwen-code-core",
-  "version": "0.7.1",
+  "version": "0.7.0",
  "description": "Qwen Code Core",
  "repository": {
    "type": "git",
--- a/packages/core/src/config/config.ts
+++ b/packages/core/src/config/config.ts
@@ -404,7 +404,7 @@ export class Config {
  private toolRegistry!: ToolRegistry;
  private promptRegistry!: PromptRegistry;
  private subagentManager!: SubagentManager;
-  private skillManager: SkillManager | null = null;
+  private skillManager!: SkillManager;
  private fileSystemService: FileSystemService;
  private contentGeneratorConfig!: ContentGeneratorConfig;
  private contentGeneratorConfigSources: ContentGeneratorConfigSources = {};
@@ -672,10 +672,8 @@ export class Config {
    }
    this.promptRegistry = new PromptRegistry();
    this.subagentManager = new SubagentManager(this);
-    if (this.getExperimentalSkills()) {
-      this.skillManager = new SkillManager(this);
-      await this.skillManager.startWatching();
-    }
+    this.skillManager = new SkillManager(this);
+    await this.skillManager.startWatching();

    // Load session subagents if they were provided before initialization
    if (this.sessionSubagents.length > 0) {
@@ -708,15 +706,12 @@ export class Config {
   * Exclusive for `OpenAIKeyPrompt` to update credentials via `/auth`
   * Delegates to ModelsConfig.
   */
-  updateCredentials(
-    credentials: {
-      apiKey?: string;
-      baseUrl?: string;
-      model?: string;
-    },
-    settingsGenerationConfig?: Partial<ContentGeneratorConfig>,
-  ): void {
-    this._modelsConfig.updateCredentials(credentials, settingsGenerationConfig);
+  updateCredentials(credentials: {
+    apiKey?: string;
+    baseUrl?: string;
+    model?: string;
+  }): void {
+    this._modelsConfig.updateCredentials(credentials);
  }

  /**
@@ -1444,7 +1439,7 @@ export class Config {
    return this.subagentManager;
  }

-  getSkillManager(): SkillManager | null {
+  getSkillManager(): SkillManager {
    return this.skillManager;
  }

--- a/packages/core/src/core/anthropicContentGenerator/converter.test.ts
+++ b/packages/core/src/core/anthropicContentGenerator/converter.test.ts
@@ -208,238 +208,6 @@ describe('AnthropicContentConverter', () => {
        ],
      });
    });
-
-    it('converts function response with inlineData image parts into tool_result with images', () => {
-      const { messages } = converter.convertGeminiRequestToAnthropic({
-        model: 'models/test',
-        contents: [
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call-1',
-                  name: 'Read',
-                  response: { output: 'Image content' },
-                  parts: [
-                    {
-                      inlineData: {
-                        mimeType: 'image/png',
-                        data: 'base64encodeddata',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      });
-
-      expect(messages).toEqual([
-        {
-          role: 'user',
-          content: [
-            {
-              type: 'tool_result',
-              tool_use_id: 'call-1',
-              content: [
-                { type: 'text', text: 'Image content' },
-                {
-                  type: 'image',
-                  source: {
-                    type: 'base64',
-                    media_type: 'image/png',
-                    data: 'base64encodeddata',
-                  },
-                },
-              ],
-            },
-          ],
-        },
-      ]);
-    });
-
-    it('renders non-image inlineData as a text block (avoids invalid image media_type)', () => {
-      const { messages } = converter.convertGeminiRequestToAnthropic({
-        model: 'models/test',
-        contents: [
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call-1',
-                  name: 'Read',
-                  response: { output: 'Audio content' },
-                  parts: [
-                    {
-                      inlineData: {
-                        mimeType: 'audio/mpeg',
-                        data: 'base64encodedaudiodata',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      });
-
-      expect(messages).toHaveLength(1);
-      expect(messages[0]?.role).toBe('user');
-
-      const toolResult = messages[0]?.content?.[0] as {
-        type: string;
-        content: Array<{ type: string; text?: string }>;
-      };
-      expect(toolResult.type).toBe('tool_result');
-      expect(Array.isArray(toolResult.content)).toBe(true);
-      expect(toolResult.content[0]).toEqual({
-        type: 'text',
-        text: 'Audio content',
-      });
-      expect(toolResult.content[1]?.type).toBe('text');
-      expect(toolResult.content[1]?.text).toContain(
-        'Unsupported inline media type for Anthropic',
-      );
-      expect(toolResult.content[1]?.text).toContain('audio/mpeg');
-    });
-
-    it('converts fileData with PDF into document block', () => {
-      const { messages } = converter.convertGeminiRequestToAnthropic({
-        model: 'models/test',
-        contents: [
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call-1',
-                  name: 'Read',
-                  response: { output: 'PDF content' },
-                  parts: [
-                    {
-                      fileData: {
-                        mimeType: 'application/pdf',
-                        fileUri: 'pdfbase64data',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      });
-
-      expect(messages).toEqual([
-        {
-          role: 'user',
-          content: [
-            {
-              type: 'tool_result',
-              tool_use_id: 'call-1',
-              content: [
-                { type: 'text', text: 'PDF content' },
-                {
-                  type: 'document',
-                  source: {
-                    type: 'base64',
-                    media_type: 'application/pdf',
-                    data: 'pdfbase64data',
-                  },
-                },
-              ],
-            },
-          ],
-        },
-      ]);
-    });
-
-    it('associates each image with its preceding functionResponse', () => {
-      const { messages } = converter.convertGeminiRequestToAnthropic({
-        model: 'models/test',
-        contents: [
-          {
-            role: 'user',
-            parts: [
-              // Tool 1 with image 1
-              {
-                functionResponse: {
-                  id: 'call-1',
-                  name: 'Read',
-                  response: { output: 'File 1' },
-                  parts: [
-                    {
-                      inlineData: {
-                        mimeType: 'image/png',
-                        data: 'image1data',
-                      },
-                    },
-                  ],
-                },
-              },
-              // Tool 2 with image 2
-              {
-                functionResponse: {
-                  id: 'call-2',
-                  name: 'Read',
-                  response: { output: 'File 2' },
-                  parts: [
-                    {
-                      inlineData: {
-                        mimeType: 'image/jpeg',
-                        data: 'image2data',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      });
-
-      // Multiple tool_result blocks are emitted in order
-      expect(messages).toHaveLength(1);
-      expect(messages[0]).toEqual({
-        role: 'user',
-        content: [
-          {
-            type: 'tool_result',
-            tool_use_id: 'call-1',
-            content: [
-              { type: 'text', text: 'File 1' },
-              {
-                type: 'image',
-                source: {
-                  type: 'base64',
-                  media_type: 'image/png',
-                  data: 'image1data',
-                },
-              },
-            ],
-          },
-          {
-            type: 'tool_result',
-            tool_use_id: 'call-2',
-            content: [
-              { type: 'text', text: 'File 2' },
-              {
-                type: 'image',
-                source: {
-                  type: 'base64',
-                  media_type: 'image/jpeg',
-                  data: 'image2data',
-                },
-              },
-            ],
-          },
-        ],
-      });
-    });
  });

  describe('convertGeminiToolsToAnthropic', () => {
--- a/packages/core/src/core/anthropicContentGenerator/converter.ts
+++ b/packages/core/src/core/anthropicContentGenerator/converter.ts
@@ -10,6 +10,7 @@ import type {
  Content,
  ContentListUnion,
  ContentUnion,
+  FunctionCall,
  FunctionResponse,
  GenerateContentParameters,
  Part,
@@ -29,6 +30,15 @@ type AnthropicMessageParam = Anthropic.MessageParam;
 type AnthropicToolParam = Anthropic.Tool;
 type AnthropicContentBlockParam = Anthropic.ContentBlockParam;

+type ThoughtPart = { text: string; signature?: string };
+
+interface ParsedParts {
+  thoughtParts: ThoughtPart[];
+  contentParts: string[];
+  functionCalls: FunctionCall[];
+  functionResponses: FunctionResponse[];
+}
+
 export class AnthropicContentConverter {
  private model: string;
  private schemaCompliance: SchemaComplianceMode;
@@ -218,161 +228,127 @@ export class AnthropicContentConverter {
    }

    if (!this.isContentObject(content)) return;
-    const parts = content.parts || [];
+
+    const parsed = this.parseParts(content.parts || []);
+
+    if (parsed.functionResponses.length > 0) {
+      for (const response of parsed.functionResponses) {
+        messages.push({
+          role: 'user',
+          content: [
+            {
+              type: 'tool_result',
+              tool_use_id: response.id || '',
+              content: this.extractFunctionResponseContent(response.response),
+            },
+          ],
+        });
+      }
+      return;
+    }
+
+    if (content.role === 'model' && parsed.functionCalls.length > 0) {
+      const thinkingBlocks: AnthropicContentBlockParam[] =
+        parsed.thoughtParts.map((part) => {
+          const thinkingBlock: unknown = {
+            type: 'thinking',
+            thinking: part.text,
+          };
+          if (part.signature) {
+            (thinkingBlock as { signature?: string }).signature =
+              part.signature;
+          }
+          return thinkingBlock as AnthropicContentBlockParam;
+        });
+      const toolUses: AnthropicContentBlockParam[] = parsed.functionCalls.map(
+        (call, index) => ({
+          type: 'tool_use',
+          id: call.id || `tool_${index}`,
+          name: call.name || '',
+          input: (call.args as Record<string, unknown>) || {},
+        }),
+      );
+
+      const textBlocks: AnthropicContentBlockParam[] = parsed.contentParts.map(
+        (text) => ({
+          type: 'text' as const,
+          text,
+        }),
+      );
+
+      messages.push({
+        role: 'assistant',
+        content: [...thinkingBlocks, ...textBlocks, ...toolUses],
+      });
+      return;
+    }
+
    const role = content.role === 'model' ? 'assistant' : 'user';
-    const contentBlocks: AnthropicContentBlockParam[] = [];
-    let toolCallIndex = 0;
+    const thinkingBlocks: AnthropicContentBlockParam[] =
+      role === 'assistant'
+        ? parsed.thoughtParts.map((part) => {
+            const thinkingBlock: unknown = {
+              type: 'thinking',
+              thinking: part.text,
+            };
+            if (part.signature) {
+              (thinkingBlock as { signature?: string }).signature =
+                part.signature;
+            }
+            return thinkingBlock as AnthropicContentBlockParam;
+          })
+        : [];
+    const textBlocks: AnthropicContentBlockParam[] = [
+      ...thinkingBlocks,
+      ...parsed.contentParts.map((text) => ({
+        type: 'text' as const,
+        text,
+      })),
+    ];
+    if (textBlocks.length > 0) {
+      messages.push({ role, content: textBlocks });
+    }
+  }
+
+  private parseParts(parts: Part[]): ParsedParts {
+    const thoughtParts: ThoughtPart[] = [];
+    const contentParts: string[] = [];
+    const functionCalls: FunctionCall[] = [];
+    const functionResponses: FunctionResponse[] = [];

    for (const part of parts) {
      if (typeof part === 'string') {
-        contentBlocks.push({ type: 'text', text: part });
-        continue;
-      }
-
-      if ('text' in part && 'thought' in part && part.thought) {
-        if (role === 'assistant') {
-          const thinkingBlock: unknown = {
-            type: 'thinking',
-            thinking: part.text || '',
-          };
-          if (
+        contentParts.push(part);
+      } else if (
+        'text' in part &&
+        part.text &&
+        !('thought' in part && part.thought)
+      ) {
+        contentParts.push(part.text);
+      } else if ('text' in part && 'thought' in part && part.thought) {
+        thoughtParts.push({
+          text: part.text || '',
+          signature:
            'thoughtSignature' in part &&
            typeof part.thoughtSignature === 'string'
-          ) {
-            (thinkingBlock as { signature?: string }).signature =
-              part.thoughtSignature;
-          }
-          contentBlocks.push(thinkingBlock as AnthropicContentBlockParam);
-        }
+              ? part.thoughtSignature
+              : undefined,
+        });
+      } else if ('functionCall' in part && part.functionCall) {
+        functionCalls.push(part.functionCall);
+      } else if ('functionResponse' in part && part.functionResponse) {
+        functionResponses.push(part.functionResponse);
      }
-
-      if ('text' in part && part.text && !('thought' in part && part.thought)) {
-        contentBlocks.push({ type: 'text', text: part.text });
-      }
-
-      const mediaBlock = this.createMediaBlockFromPart(part);
-      if (mediaBlock) {
-        contentBlocks.push(mediaBlock);
-      }
-
-      if ('functionCall' in part && part.functionCall) {
-        if (role === 'assistant') {
-          contentBlocks.push({
-            type: 'tool_use',
-            id: part.functionCall.id || `tool_${toolCallIndex}`,
-            name: part.functionCall.name || '',
-            input: (part.functionCall.args as Record<string, unknown>) || {},
-          });
-          toolCallIndex += 1;
-        }
-      }
-
-      if (part.functionResponse) {
-        const toolResultBlock = this.createToolResultBlock(
-          part.functionResponse,
-        );
-        if (toolResultBlock && role === 'user') {
-          contentBlocks.push(toolResultBlock);
-        }
-      }
-    }
-
-    if (contentBlocks.length > 0) {
-      messages.push({ role, content: contentBlocks });
-    }
-  }
-
-  private createToolResultBlock(
-    response: FunctionResponse,
-  ): Anthropic.ToolResultBlockParam | null {
-    const textContent = this.extractFunctionResponseContent(response.response);
-
-    type ToolResultContent = Anthropic.ToolResultBlockParam['content'];
-    const partBlocks: AnthropicContentBlockParam[] = [];
-
-    for (const part of response.parts || []) {
-      const block = this.createMediaBlockFromPart(part);
-      if (block) {
-        partBlocks.push(block);
-      }
-    }
-
-    let content: ToolResultContent;
-    if (partBlocks.length > 0) {
-      const blocks: AnthropicContentBlockParam[] = [];
-      if (textContent) {
-        blocks.push({ type: 'text', text: textContent });
-      }
-      blocks.push(...partBlocks);
-      content = blocks as unknown as ToolResultContent;
-    } else {
-      content = textContent;
    }

    return {
-      type: 'tool_result',
-      tool_use_id: response.id || '',
-      content,
+      thoughtParts,
+      contentParts,
+      functionCalls,
+      functionResponses,
    };
  }

-  private createMediaBlockFromPart(
-    part: Part,
-  ): AnthropicContentBlockParam | null {
-    if (part.inlineData?.mimeType && part.inlineData?.data) {
-      if (!this.isSupportedAnthropicImageMimeType(part.inlineData.mimeType)) {
-        const displayName = part.inlineData.displayName ?? '';
-        return {
-          type: 'text',
-          text: `Unsupported inline media type for Anthropic: ${part.inlineData.mimeType}${displayName}.`,
-        };
-      }
-      return {
-        type: 'image',
-        source: {
-          type: 'base64',
-          media_type: part.inlineData.mimeType as
-            | 'image/jpeg'
-            | 'image/png'
-            | 'image/gif'
-            | 'image/webp',
-          data: part.inlineData.data,
-        },
-      };
-    }
-
-    if (part.fileData?.mimeType && part.fileData?.fileUri) {
-      if (part.fileData.mimeType !== 'application/pdf') {
-        const displayName = part.fileData.displayName ?? '';
-        return {
-          type: 'text',
-          text: `Unsupported file media for Anthropic: ${part.fileData.mimeType}${displayName}`,
-        };
-      }
-      return {
-        type: 'document',
-        source: {
-          type: 'base64',
-          media_type: part.fileData.mimeType as 'application/pdf',
-          data: part.fileData.fileUri,
-        },
-      };
-    }
-
-    return null;
-  }
-
-  private isSupportedAnthropicImageMimeType(
-    mimeType: string,
-  ): mimeType is 'image/jpeg' | 'image/png' | 'image/gif' | 'image/webp' {
-    return (
-      mimeType === 'image/jpeg' ||
-      mimeType === 'image/png' ||
-      mimeType === 'image/gif' ||
-      mimeType === 'image/webp'
-    );
-  }
-
  private extractTextFromContentUnion(contentUnion: unknown): string {
    if (typeof contentUnion === 'string') {
      return contentUnion;
--- a/packages/core/src/core/contentGenerator.ts
+++ b/packages/core/src/core/contentGenerator.ts
@@ -270,28 +270,28 @@ export function createContentGeneratorConfig(
 }

 export async function createContentGenerator(
-  generatorConfig: ContentGeneratorConfig,
-  config: Config,
+  config: ContentGeneratorConfig,
+  gcConfig: Config,
  isInitialAuth?: boolean,
 ): Promise<ContentGenerator> {
-  const validation = validateModelConfig(generatorConfig, false);
+  const validation = validateModelConfig(config, false);
  if (!validation.valid) {
    throw new Error(validation.errors.map((e) => e.message).join('\n'));
  }

-  const authType = generatorConfig.authType;
-  if (!authType) {
-    throw new Error('ContentGeneratorConfig must have an authType');
-  }
-
-  let baseGenerator: ContentGenerator;
-
-  if (authType === AuthType.USE_OPENAI) {
+  if (config.authType === AuthType.USE_OPENAI) {
+    // Import OpenAIContentGenerator dynamically to avoid circular dependencies
    const { createOpenAIContentGenerator } = await import(
      './openaiContentGenerator/index.js'
    );
-    baseGenerator = createOpenAIContentGenerator(generatorConfig, config);
-  } else if (authType === AuthType.QWEN_OAUTH) {
+
+    // Always use OpenAIContentGenerator, logging is controlled by enableOpenAILogging flag
+    const generator = createOpenAIContentGenerator(config, gcConfig);
+    return new LoggingContentGenerator(generator, gcConfig);
+  }
+
+  if (config.authType === AuthType.QWEN_OAUTH) {
+    // Import required classes dynamically
    const { getQwenOAuthClient: getQwenOauthClient } = await import(
      '../qwen/qwenOAuth2.js'
    );
@@ -300,38 +300,44 @@ export async function createContentGenerator(
    );

    try {
+      // Get the Qwen OAuth client (now includes integrated token management)
+      // If this is initial auth, require cached credentials to detect missing credentials
      const qwenClient = await getQwenOauthClient(
-        config,
+        gcConfig,
        isInitialAuth ? { requireCachedCredentials: true } : undefined,
      );
-      baseGenerator = new QwenContentGenerator(
-        qwenClient,
-        generatorConfig,
-        config,
-      );
+
+      // Create the content generator with dynamic token management
+      const generator = new QwenContentGenerator(qwenClient, config, gcConfig);
+      return new LoggingContentGenerator(generator, gcConfig);
    } catch (error) {
      throw new Error(
        `${error instanceof Error ? error.message : String(error)}`,
      );
    }
-  } else if (authType === AuthType.USE_ANTHROPIC) {
+  }
+
+  if (config.authType === AuthType.USE_ANTHROPIC) {
    const { createAnthropicContentGenerator } = await import(
      './anthropicContentGenerator/index.js'
    );
-    baseGenerator = createAnthropicContentGenerator(generatorConfig, config);
-  } else if (
-    authType === AuthType.USE_GEMINI ||
-    authType === AuthType.USE_VERTEX_AI
+
+    const generator = createAnthropicContentGenerator(config, gcConfig);
+    return new LoggingContentGenerator(generator, gcConfig);
+  }
+
+  if (
+    config.authType === AuthType.USE_GEMINI ||
+    config.authType === AuthType.USE_VERTEX_AI
  ) {
    const { createGeminiContentGenerator } = await import(
      './geminiContentGenerator/index.js'
    );
-    baseGenerator = createGeminiContentGenerator(generatorConfig, config);
-  } else {
-    throw new Error(
-      `Error creating contentGenerator: Unsupported authType: ${authType}`,
-    );
+    const generator = createGeminiContentGenerator(config, gcConfig);
+    return new LoggingContentGenerator(generator, gcConfig);
  }

-  return new LoggingContentGenerator(baseGenerator, config, generatorConfig);
+  throw new Error(
+    `Error creating contentGenerator: Unsupported authType: ${config.authType}`,
+  );
 }
--- a/packages/core/src/core/coreToolScheduler.test.ts
+++ b/packages/core/src/core/coreToolScheduler.test.ts
@@ -800,11 +800,11 @@ describe('convertToFunctionResponse', () => {
          name: toolName,
          id: callId,
          response: {
-            output: '',
+            output: 'Binary content of type image/png was processed.',
          },
-          parts: [{ inlineData: { mimeType: 'image/png', data: 'base64...' } }],
        },
      },
+      llmContent,
    ]);
  });

@@ -819,15 +819,11 @@ describe('convertToFunctionResponse', () => {
          name: toolName,
          id: callId,
          response: {
-            output: '',
+            output: 'Binary content of type application/pdf was processed.',
          },
-          parts: [
-            {
-              fileData: { mimeType: 'application/pdf', fileUri: 'gs://...' },
-            },
-          ],
        },
      },
+      llmContent,
    ]);
  });

@@ -861,13 +857,11 @@ describe('convertToFunctionResponse', () => {
          name: toolName,
          id: callId,
          response: {
-            output: '',
+            output: 'Binary content of type image/gif was processed.',
          },
-          parts: [
-            { inlineData: { mimeType: 'image/gif', data: 'gifdata...' } },
-          ],
        },
      },
+      ...llmContent,
    ]);
  });

--- a/packages/core/src/core/coreToolScheduler.ts
+++ b/packages/core/src/core/coreToolScheduler.ts
@@ -30,12 +30,7 @@ import {
  ToolOutputTruncatedEvent,
  InputFormat,
 } from '../index.js';
-import type {
-  FunctionResponse,
-  FunctionResponsePart,
-  Part,
-  PartListUnion,
-} from '@google/genai';
+import type { Part, PartListUnion } from '@google/genai';
 import { getResponseTextFromParts } from '../utils/generateContentResponseUtilities.js';
 import type { ModifyContext } from '../tools/modifiable-tool.js';
 import {
@@ -156,17 +151,13 @@ function createFunctionResponsePart(
  callId: string,
  toolName: string,
  output: string,
-  mediaParts?: FunctionResponsePart[],
 ): Part {
-  const functionResponse: FunctionResponse = {
-    id: callId,
-    name: toolName,
-    response: { output },
-    ...(mediaParts && mediaParts.length > 0 ? { parts: mediaParts } : {}),
-  };
-
  return {
-    functionResponse,
+    functionResponse: {
+      id: callId,
+      name: toolName,
+      response: { output },
+    },
  };
 }

@@ -207,21 +198,16 @@ export function convertToFunctionResponse(
  }

  if (contentToProcess.inlineData || contentToProcess.fileData) {
-    const mediaParts: FunctionResponsePart[] = [];
-    if (contentToProcess.inlineData) {
-      mediaParts.push({ inlineData: contentToProcess.inlineData });
-    }
-    if (contentToProcess.fileData) {
-      mediaParts.push({ fileData: contentToProcess.fileData });
-    }
-
+    const mimeType =
+      contentToProcess.inlineData?.mimeType ||
+      contentToProcess.fileData?.mimeType ||
+      'unknown';
    const functionResponse = createFunctionResponsePart(
      callId,
      toolName,
-      '',
-      mediaParts,
+      `Binary content of type ${mimeType} was processed.`,
    );
-    return [functionResponse];
+    return [functionResponse, contentToProcess];
  }

  if (contentToProcess.text !== undefined) {
--- a/packages/core/src/core/loggingContentGenerator/loggingContentGenerator.test.ts
+++ b/packages/core/src/core/loggingContentGenerator/loggingContentGenerator.test.ts
@@ -12,7 +12,6 @@ import type {
 import { GenerateContentResponse } from '@google/genai';
 import type { Config } from '../../config/config.js';
 import type { ContentGenerator } from '../contentGenerator.js';
-import { AuthType } from '../contentGenerator.js';
 import { LoggingContentGenerator } from './index.js';
 import { OpenAIContentConverter } from '../openaiContentGenerator/converter.js';
 import {
@@ -51,17 +50,14 @@ const convertGeminiResponseToOpenAISpy = vi
    choices: [],
  } as OpenAI.Chat.ChatCompletion);

-const createConfig = (overrides: Record<string, unknown> = {}): Config => {
-  const configContent = {
-    authType: 'openai',
-    enableOpenAILogging: false,
-    ...overrides,
-  };
-  return {
-    getContentGeneratorConfig: () => configContent,
-    getAuthType: () => configContent.authType as AuthType | undefined,
-  } as Config;
-};
+const createConfig = (overrides: Record<string, unknown> = {}): Config =>
+  ({
+    getContentGeneratorConfig: () => ({
+      authType: 'openai',
+      enableOpenAILogging: false,
+      ...overrides,
+    }),
+  }) as Config;

 const createWrappedGenerator = (
  generateContent: ContentGenerator['generateContent'],
@@ -128,17 +124,13 @@ describe('LoggingContentGenerator', () => {
      ),
      vi.fn(),
    );
-    const generatorConfig = {
-      model: 'test-model',
-      authType: AuthType.USE_OPENAI,
-      enableOpenAILogging: true,
-      openAILoggingDir: 'logs',
-      schemaCompliance: 'openapi_30' as const,
-    };
    const generator = new LoggingContentGenerator(
      wrapped,
-      createConfig(),
-      generatorConfig,
+      createConfig({
+        enableOpenAILogging: true,
+        openAILoggingDir: 'logs',
+        schemaCompliance: 'openapi_30',
+      }),
    );

    const request = {
@@ -233,15 +225,9 @@ describe('LoggingContentGenerator', () => {
      vi.fn().mockRejectedValue(error),
      vi.fn(),
    );
-    const generatorConfig = {
-      model: 'test-model',
-      authType: AuthType.USE_OPENAI,
-      enableOpenAILogging: true,
-    };
    const generator = new LoggingContentGenerator(
      wrapped,
-      createConfig(),
-      generatorConfig,
+      createConfig({ enableOpenAILogging: true }),
    );

    const request = {
@@ -307,15 +293,9 @@ describe('LoggingContentGenerator', () => {
        })(),
      ),
    );
-    const generatorConfig = {
-      model: 'test-model',
-      authType: AuthType.USE_OPENAI,
-      enableOpenAILogging: true,
-    };
    const generator = new LoggingContentGenerator(
      wrapped,
-      createConfig(),
-      generatorConfig,
+      createConfig({ enableOpenAILogging: true }),
    );

    const request = {
@@ -365,15 +345,9 @@ describe('LoggingContentGenerator', () => {
        })(),
      ),
    );
-    const generatorConfig = {
-      model: 'test-model',
-      authType: AuthType.USE_OPENAI,
-      enableOpenAILogging: true,
-    };
    const generator = new LoggingContentGenerator(
      wrapped,
-      createConfig(),
-      generatorConfig,
+      createConfig({ enableOpenAILogging: true }),
    );

    const request = {
--- a/packages/core/src/core/loggingContentGenerator/loggingContentGenerator.ts
+++ b/packages/core/src/core/loggingContentGenerator/loggingContentGenerator.ts
@@ -31,10 +31,7 @@ import {
  logApiRequest,
  logApiResponse,
 } from '../../telemetry/loggers.js';
-import type {
-  ContentGenerator,
-  ContentGeneratorConfig,
-} from '../contentGenerator.js';
+import type { ContentGenerator } from '../contentGenerator.js';
 import { isStructuredError } from '../../utils/quotaErrorDetection.js';
 import { OpenAIContentConverter } from '../openaiContentGenerator/converter.js';
 import { OpenAILogger } from '../../utils/openaiLogger.js';
@@ -53,11 +50,9 @@ export class LoggingContentGenerator implements ContentGenerator {
  constructor(
    private readonly wrapped: ContentGenerator,
    private readonly config: Config,
-    generatorConfig: ContentGeneratorConfig,
  ) {
-    // Extract fields needed for initialization from passed config
-    // (config.getContentGeneratorConfig() may not be available yet during refreshAuth)
-    if (generatorConfig.enableOpenAILogging) {
+    const generatorConfig = this.config.getContentGeneratorConfig();
+    if (generatorConfig?.enableOpenAILogging) {
      this.openaiLogger = new OpenAILogger(generatorConfig.openAILoggingDir);
      this.schemaCompliance = generatorConfig.schemaCompliance;
    }
@@ -94,7 +89,7 @@ export class LoggingContentGenerator implements ContentGenerator {
        model,
        durationMs,
        prompt_id,
-        this.config.getAuthType(),
+        this.config.getContentGeneratorConfig()?.authType,
        usageMetadata,
        responseText,
      ),
@@ -131,7 +126,7 @@ export class LoggingContentGenerator implements ContentGenerator {
        errorMessage,
        durationMs,
        prompt_id,
-        this.config.getAuthType(),
+        this.config.getContentGeneratorConfig()?.authType,
        errorType,
        errorStatus,
      ),
--- a/packages/core/src/core/nonInteractiveToolExecutor.test.ts
+++ b/packages/core/src/core/nonInteractiveToolExecutor.test.ts
@@ -309,13 +309,11 @@ describe('executeToolCall', () => {
            name: 'testTool',
            id: 'call6',
            response: {
-              output: '',
+              output: 'Binary content of type image/png was processed.',
            },
-            parts: [
-              { inlineData: { mimeType: 'image/png', data: 'base64data' } },
-            ],
          },
        },
+        imageDataPart,
      ],
    });
  });
--- a/packages/core/src/core/openaiContentGenerator/converter.test.ts
+++ b/packages/core/src/core/openaiContentGenerator/converter.test.ts
@@ -122,13 +122,7 @@ describe('OpenAIContentConverter', () => {
      const toolMessage = messages.find((message) => message.role === 'tool');

      expect(toolMessage).toBeDefined();
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-      }>;
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('Raw output text');
+      expect(toolMessage?.content).toBe('Raw output text');
    });

    it('should prioritize error field when present', () => {
@@ -140,13 +134,7 @@ describe('OpenAIContentConverter', () => {
      const toolMessage = messages.find((message) => message.role === 'tool');

      expect(toolMessage).toBeDefined();
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-      }>;
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('Command failed');
+      expect(toolMessage?.content).toBe('Command failed');
    });

    it('should stringify non-string responses', () => {
@@ -158,318 +146,7 @@ describe('OpenAIContentConverter', () => {
      const toolMessage = messages.find((message) => message.role === 'tool');

      expect(toolMessage).toBeDefined();
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-      }>;
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('{"data":{"value":42}}');
-    });
-
-    it('should convert function responses with inlineData to tool message with embedded image_url', () => {
-      const request: GenerateContentParameters = {
-        model: 'models/test',
-        contents: [
-          {
-            role: 'model',
-            parts: [
-              {
-                functionCall: {
-                  id: 'call_1',
-                  name: 'Read',
-                  args: {},
-                },
-              },
-            ],
-          },
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call_1',
-                  name: 'Read',
-                  response: { output: 'Image content' },
-                  parts: [
-                    {
-                      inlineData: {
-                        mimeType: 'image/png',
-                        data: 'base64encodedimagedata',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      };
-
-      const messages = converter.convertGeminiRequestToOpenAI(request);
-
-      // Should have tool message with both text and image content
-      const toolMessage = messages.find((message) => message.role === 'tool');
-      expect(toolMessage).toBeDefined();
-      expect((toolMessage as { tool_call_id?: string }).tool_call_id).toBe(
-        'call_1',
-      );
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-        image_url?: { url: string };
-      }>;
-      expect(contentArray).toHaveLength(2);
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('Image content');
-      expect(contentArray[1].type).toBe('image_url');
-      expect(contentArray[1].image_url?.url).toBe(
-        'data:image/png;base64,base64encodedimagedata',
-      );
-
-      // No separate user message should be created
-      const userMessage = messages.find((message) => message.role === 'user');
-      expect(userMessage).toBeUndefined();
-    });
-
-    it('should convert function responses with fileData to tool message with embedded input_file', () => {
-      const request: GenerateContentParameters = {
-        model: 'models/test',
-        contents: [
-          {
-            role: 'model',
-            parts: [
-              {
-                functionCall: {
-                  id: 'call_1',
-                  name: 'Read',
-                  args: {},
-                },
-              },
-            ],
-          },
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call_1',
-                  name: 'Read',
-                  response: { output: 'File content' },
-                  parts: [
-                    {
-                      fileData: {
-                        mimeType: 'image/jpeg',
-                        fileUri: 'base64imagedata',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      };
-
-      const messages = converter.convertGeminiRequestToOpenAI(request);
-
-      // Should have tool message with both text and file content
-      const toolMessage = messages.find((message) => message.role === 'tool');
-      expect(toolMessage).toBeDefined();
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-        file?: { filename: string; file_data: string };
-      }>;
-      expect(contentArray).toHaveLength(2);
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('File content');
-      expect(contentArray[1].type).toBe('file');
-      expect(contentArray[1].file?.filename).toBe('file'); // Default filename when displayName not provided
-      expect(contentArray[1].file?.file_data).toBe(
-        'data:image/jpeg;base64,base64imagedata',
-      );
-
-      // No separate user message should be created
-      const userMessage = messages.find((message) => message.role === 'user');
-      expect(userMessage).toBeUndefined();
-    });
-
-    it('should convert PDF fileData to tool message with embedded input_file', () => {
-      const request: GenerateContentParameters = {
-        model: 'models/test',
-        contents: [
-          {
-            role: 'model',
-            parts: [
-              {
-                functionCall: {
-                  id: 'call_1',
-                  name: 'Read',
-                  args: {},
-                },
-              },
-            ],
-          },
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call_1',
-                  name: 'Read',
-                  response: { output: 'PDF content' },
-                  parts: [
-                    {
-                      fileData: {
-                        mimeType: 'application/pdf',
-                        fileUri: 'base64pdfdata',
-                        displayName: 'document.pdf',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      };
-
-      const messages = converter.convertGeminiRequestToOpenAI(request);
-
-      // Should have tool message with both text and file content
-      const toolMessage = messages.find((message) => message.role === 'tool');
-      expect(toolMessage).toBeDefined();
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-        file?: { filename: string; file_data: string };
-      }>;
-      expect(contentArray).toHaveLength(2);
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('PDF content');
-      expect(contentArray[1].type).toBe('file');
-      expect(contentArray[1].file?.filename).toBe('document.pdf');
-      expect(contentArray[1].file?.file_data).toBe(
-        'data:application/pdf;base64,base64pdfdata',
-      );
-
-      // No separate user message should be created
-      const userMessage = messages.find((message) => message.role === 'user');
-      expect(userMessage).toBeUndefined();
-    });
-
-    it('should convert audio parts to tool message with embedded input_audio', () => {
-      const request: GenerateContentParameters = {
-        model: 'models/test',
-        contents: [
-          {
-            role: 'model',
-            parts: [
-              {
-                functionCall: {
-                  id: 'call_1',
-                  name: 'Record',
-                  args: {},
-                },
-              },
-            ],
-          },
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call_1',
-                  name: 'Record',
-                  response: { output: 'Audio recorded' },
-                  parts: [
-                    {
-                      inlineData: {
-                        mimeType: 'audio/wav',
-                        data: 'audiobase64data',
-                      },
-                    },
-                  ],
-                },
-              },
-            ],
-          },
-        ],
-      };
-
-      const messages = converter.convertGeminiRequestToOpenAI(request);
-
-      // Should have tool message with both text and audio content
-      const toolMessage = messages.find((message) => message.role === 'tool');
-      expect(toolMessage).toBeDefined();
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-        input_audio?: { data: string; format: string };
-      }>;
-      expect(contentArray).toHaveLength(2);
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('Audio recorded');
-      expect(contentArray[1].type).toBe('input_audio');
-      expect(contentArray[1].input_audio?.data).toBe('audiobase64data');
-      expect(contentArray[1].input_audio?.format).toBe('wav');
-
-      // No separate user message should be created
-      const userMessage = messages.find((message) => message.role === 'user');
-      expect(userMessage).toBeUndefined();
-    });
-
-    it('should create tool message with text-only content when no media parts', () => {
-      const request = createRequestWithFunctionResponse({
-        output: 'Plain text output',
-      });
-
-      const messages = converter.convertGeminiRequestToOpenAI(request);
-      const toolMessage = messages.find((message) => message.role === 'tool');
-
-      expect(toolMessage).toBeDefined();
-      expect(Array.isArray(toolMessage?.content)).toBe(true);
-      const contentArray = toolMessage?.content as Array<{
-        type: string;
-        text?: string;
-      }>;
-      expect(contentArray).toHaveLength(1);
-      expect(contentArray[0].type).toBe('text');
-      expect(contentArray[0].text).toBe('Plain text output');
-
-      // No user message should be created when there's no media
-      const userMessage = messages.find((message) => message.role === 'user');
-      expect(userMessage).toBeUndefined();
-    });
-
-    it('should skip empty function responses with no media and no text', () => {
-      const request: GenerateContentParameters = {
-        model: 'models/test',
-        contents: [
-          {
-            role: 'user',
-            parts: [
-              {
-                functionResponse: {
-                  id: 'call_1',
-                  name: 'Empty',
-                  response: { output: '' },
-                },
-              },
-            ],
-          },
-        ],
-      };
-
-      const messages = converter.convertGeminiRequestToOpenAI(request);
-
-      // Should have no messages for empty response
-      expect(messages).toHaveLength(0);
+      expect(toolMessage?.content).toBe('{"data":{"value":42}}');
    });
  });

@@ -503,35 +180,6 @@ describe('OpenAIContentConverter', () => {
      );
    });

-    it('should convert reasoning to a thought part for non-streaming responses', () => {
-      const response = converter.convertOpenAIResponseToGemini({
-        object: 'chat.completion',
-        id: 'chatcmpl-2',
-        created: 123,
-        model: 'gpt-test',
-        choices: [
-          {
-            index: 0,
-            message: {
-              role: 'assistant',
-              content: 'final answer',
-              reasoning: 'chain-of-thought',
-            },
-            finish_reason: 'stop',
-            logprobs: null,
-          },
-        ],
-      } as unknown as OpenAI.Chat.ChatCompletion);
-
-      const parts = response.candidates?.[0]?.content?.parts;
-      expect(parts?.[0]).toEqual(
-        expect.objectContaining({ thought: true, text: 'chain-of-thought' }),
-      );
-      expect(parts?.[1]).toEqual(
-        expect.objectContaining({ text: 'final answer' }),
-      );
-    });
-
    it('should convert streaming reasoning_content delta to a thought part', () => {
      const chunk = converter.convertOpenAIChunkToGemini({
        object: 'chat.completion.chunk',
@@ -560,34 +208,6 @@ describe('OpenAIContentConverter', () => {
      );
    });

-    it('should convert streaming reasoning delta to a thought part', () => {
-      const chunk = converter.convertOpenAIChunkToGemini({
-        object: 'chat.completion.chunk',
-        id: 'chunk-1b',
-        created: 456,
-        choices: [
-          {
-            index: 0,
-            delta: {
-              content: 'visible text',
-              reasoning: 'thinking...',
-            },
-            finish_reason: 'stop',
-            logprobs: null,
-          },
-        ],
-        model: 'gpt-test',
-      } as unknown as OpenAI.Chat.ChatCompletionChunk);
-
-      const parts = chunk.candidates?.[0]?.content?.parts;
-      expect(parts?.[0]).toEqual(
-        expect.objectContaining({ thought: true, text: 'thinking...' }),
-      );
-      expect(parts?.[1]).toEqual(
-        expect.objectContaining({ text: 'visible text' }),
-      );
-    });
-
    it('should not throw when streaming chunk has no delta', () => {
      const chunk = converter.convertOpenAIChunkToGemini({
        object: 'chat.completion.chunk',
@@ -964,7 +584,11 @@ describe('OpenAIContentConverter', () => {

      expect(messages).toHaveLength(1);
      expect(messages[0].role).toBe('assistant');
-      expect(messages[0].content).toBe('First partSecond part');
+      const content = messages[0]
+        .content as OpenAI.Chat.ChatCompletionContentPart[];
+      expect(content).toHaveLength(2);
+      expect(content[0]).toEqual({ type: 'text', text: 'First part' });
+      expect(content[1]).toEqual({ type: 'text', text: 'Second part' });
    });

    it('should merge multiple consecutive assistant messages', () => {
@@ -990,7 +614,9 @@ describe('OpenAIContentConverter', () => {

      expect(messages).toHaveLength(1);
      expect(messages[0].role).toBe('assistant');
-      expect(messages[0].content).toBe('Part 1Part 2Part 3');
+      const content = messages[0]
+        .content as OpenAI.Chat.ChatCompletionContentPart[];
+      expect(content).toHaveLength(3);
    });

    it('should merge tool_calls from consecutive assistant messages', () => {
@@ -1048,9 +674,7 @@ describe('OpenAIContentConverter', () => {
        ],
      };

-      const messages = converter.convertGeminiRequestToOpenAI(request, {
-        cleanOrphanToolCalls: false,
-      });
+      const messages = converter.convertGeminiRequestToOpenAI(request);

      // Should have: assistant (tool_call_1), tool (result_1), assistant (tool_call_2), tool (result_2)
      expect(messages).toHaveLength(4);
@@ -1105,7 +729,10 @@ describe('OpenAIContentConverter', () => {
      const messages = converter.convertGeminiRequestToOpenAI(request);

      expect(messages).toHaveLength(1);
-      expect(messages[0].content).toBe('Text partAnother text');
+      const content = messages[0]
+        .content as OpenAI.Chat.ChatCompletionContentPart[];
+      expect(Array.isArray(content)).toBe(true);
+      expect(content).toHaveLength(2);
    });

    it('should merge empty content correctly', () => {
@@ -1131,7 +758,11 @@ describe('OpenAIContentConverter', () => {

      // Empty messages should be filtered out
      expect(messages).toHaveLength(1);
-      expect(messages[0].content).toBe('FirstSecond');
+      const content = messages[0]
+        .content as OpenAI.Chat.ChatCompletionContentPart[];
+      expect(content).toHaveLength(2);
+      expect(content[0]).toEqual({ type: 'text', text: 'First' });
+      expect(content[1]).toEqual({ type: 'text', text: 'Second' });
    });
  });
 });
--- a/packages/core/src/core/openaiContentGenerator/converter.ts
+++ b/packages/core/src/core/openaiContentGenerator/converter.ts
@@ -11,6 +11,7 @@ import type {
  Tool,
  ToolListUnion,
  CallableTool,
+  FunctionCall,
  FunctionResponse,
  ContentListUnion,
  ContentUnion,
@@ -46,13 +47,11 @@ type ExtendedChatCompletionMessageParam =
 export interface ExtendedCompletionMessage
  extends OpenAI.Chat.ChatCompletionMessage {
  reasoning_content?: string | null;
-  reasoning?: string | null;
 }

 export interface ExtendedCompletionChunkDelta
  extends OpenAI.Chat.ChatCompletionChunk.Choice.Delta {
  reasoning_content?: string | null;
-  reasoning?: string | null;
 }

 /**
@@ -64,17 +63,21 @@ export interface ToolCallAccumulator {
  arguments: string;
 }

-type OpenAIContentPart =
-  | OpenAI.Chat.ChatCompletionContentPartText
-  | OpenAI.Chat.ChatCompletionContentPartImage
-  | OpenAI.Chat.ChatCompletionContentPartInputAudio
-  | {
-      type: 'file';
-      file: {
-        filename: string;
-        file_data: string;
-      };
-    };
+/**
+ * Parsed parts from Gemini content, categorized by type
+ */
+interface ParsedParts {
+  thoughtParts: string[];
+  contentParts: string[];
+  functionCalls: FunctionCall[];
+  functionResponses: FunctionResponse[];
+  mediaParts: Array<{
+    type: 'image' | 'audio' | 'file';
+    data: string;
+    mimeType: string;
+    fileUri?: string;
+  }>;
+}

 /**
 * Converter class for transforming data between Gemini and OpenAI formats
@@ -268,48 +271,28 @@ export class OpenAIContentConverter {
  ): OpenAI.Chat.ChatCompletion {
    const candidate = response.candidates?.[0];
    const parts = (candidate?.content?.parts || []) as Part[];
-
-    // Parse parts inline
-    const thoughtParts: string[] = [];
-    const contentParts: string[] = [];
-    const toolCalls: OpenAI.Chat.ChatCompletionMessageToolCall[] = [];
-    let toolCallIndex = 0;
-
-    for (const part of parts) {
-      if (typeof part === 'string') {
-        contentParts.push(part);
-      } else if ('text' in part && part.text) {
-        if ('thought' in part && part.thought) {
-          thoughtParts.push(part.text);
-        } else {
-          contentParts.push(part.text);
-        }
-      } else if ('functionCall' in part && part.functionCall) {
-        toolCalls.push({
-          id: part.functionCall.id || `call_${toolCallIndex}`,
-          type: 'function' as const,
-          function: {
-            name: part.functionCall.name || '',
-            arguments: JSON.stringify(part.functionCall.args || {}),
-          },
-        });
-        toolCallIndex += 1;
-      }
-    }
+    const parsedParts = this.parseParts(parts);

    const message: ExtendedCompletionMessage = {
      role: 'assistant',
-      content: contentParts.join('') || null,
+      content: parsedParts.contentParts.join('') || null,
      refusal: null,
    };

-    const reasoningContent = thoughtParts.join('');
+    const reasoningContent = parsedParts.thoughtParts.join('');
    if (reasoningContent) {
      message.reasoning_content = reasoningContent;
    }

-    if (toolCalls.length > 0) {
-      message.tool_calls = toolCalls;
+    if (parsedParts.functionCalls.length > 0) {
+      message.tool_calls = parsedParts.functionCalls.map((call, index) => ({
+        id: call.id || `call_${index}`,
+        type: 'function' as const,
+        function: {
+          name: call.name || '',
+          arguments: JSON.stringify(call.args || {}),
+        },
+      }));
    }

    const finishReason = this.mapGeminiFinishReasonToOpenAI(
@@ -407,82 +390,40 @@ export class OpenAIContentConverter {
    }

    if (!this.isContentObject(content)) return;
-    const parts = content.parts || [];
-    const role = content.role === 'model' ? 'assistant' : 'user';

-    const contentParts: OpenAIContentPart[] = [];
-    const reasoningParts: string[] = [];
-    const toolCalls: OpenAI.Chat.ChatCompletionMessageToolCall[] = [];
-    let toolCallIndex = 0;
+    const parsedParts = this.parseParts(content.parts || []);

-    for (const part of parts) {
-      if (typeof part === 'string') {
-        contentParts.push({ type: 'text' as const, text: part });
-        continue;
-      }
-
-      if ('text' in part && 'thought' in part && part.thought) {
-        if (role === 'assistant' && part.text) {
-          reasoningParts.push(part.text);
-        }
-      }
-
-      if ('text' in part && part.text && !('thought' in part && part.thought)) {
-        contentParts.push({ type: 'text' as const, text: part.text });
-      }
-
-      const mediaPart = this.createMediaContentPart(part);
-      if (mediaPart && role === 'user') {
-        contentParts.push(mediaPart);
-      }
-
-      if ('functionCall' in part && part.functionCall && role === 'assistant') {
-        toolCalls.push({
-          id: part.functionCall.id || `call_${toolCallIndex}`,
-          type: 'function' as const,
-          function: {
-            name: part.functionCall.name || '',
-            arguments: JSON.stringify(part.functionCall.args || {}),
-          },
+    // Handle function responses (tool results) first
+    if (parsedParts.functionResponses.length > 0) {
+      for (const funcResponse of parsedParts.functionResponses) {
+        messages.push({
+          role: 'tool' as const,
+          tool_call_id: funcResponse.id || '',
+          content: this.extractFunctionResponseContent(funcResponse.response),
        });
-        toolCallIndex += 1;
-      }
-
-      if (part.functionResponse && role === 'user') {
-        // Create tool message for the function response (with embedded media)
-        const toolMessage = this.createToolMessage(part.functionResponse);
-        if (toolMessage) {
-          messages.push(toolMessage);
-        }
      }
+      return;
    }

-    if (role === 'assistant') {
-      if (
-        contentParts.length === 0 &&
-        toolCalls.length === 0 &&
-        reasoningParts.length === 0
-      ) {
-        return;
-      }
+    // Handle model messages with function calls
+    if (content.role === 'model' && parsedParts.functionCalls.length > 0) {
+      const toolCalls = parsedParts.functionCalls.map((fc, index) => ({
+        id: fc.id || `call_${index}`,
+        type: 'function' as const,
+        function: {
+          name: fc.name || '',
+          arguments: JSON.stringify(fc.args || {}),
+        },
+      }));

-      const assistantTextContent = contentParts
-        .filter(
-          (part): part is OpenAI.Chat.ChatCompletionContentPartText =>
-            part.type === 'text',
-        )
-        .map((part) => part.text)
-        .join('');
      const assistantMessage: ExtendedChatCompletionAssistantMessageParam = {
-        role: 'assistant',
-        content: assistantTextContent || null,
+        role: 'assistant' as const,
+        content: parsedParts.contentParts.join('') || null,
+        tool_calls: toolCalls,
      };

-      if (toolCalls.length > 0) {
-        assistantMessage.tool_calls = toolCalls;
-      }
-
-      const reasoningContent = reasoningParts.join('');
+      // Only include reasoning_content if it has actual content
+      const reasoningContent = parsedParts.thoughtParts.join('');
      if (reasoningContent) {
        assistantMessage.reasoning_content = reasoningContent;
      }
@@ -491,15 +432,79 @@ export class OpenAIContentConverter {
      return;
    }

-    if (contentParts.length > 0) {
-      messages.push({
-        role: 'user',
-        content:
-          contentParts as unknown as OpenAI.Chat.ChatCompletionContentPart[],
-      });
+    // Handle regular messages with multimodal content
+    const role = content.role === 'model' ? 'assistant' : 'user';
+    const openAIMessage = this.createMultimodalMessage(role, parsedParts);
+
+    if (openAIMessage) {
+      messages.push(openAIMessage);
    }
  }

+  /**
+   * Parse Gemini parts into categorized components
+   */
+  private parseParts(parts: Part[]): ParsedParts {
+    const thoughtParts: string[] = [];
+    const contentParts: string[] = [];
+    const functionCalls: FunctionCall[] = [];
+    const functionResponses: FunctionResponse[] = [];
+    const mediaParts: Array<{
+      type: 'image' | 'audio' | 'file';
+      data: string;
+      mimeType: string;
+      fileUri?: string;
+    }> = [];
+
+    for (const part of parts) {
+      if (typeof part === 'string') {
+        contentParts.push(part);
+      } else if (
+        'text' in part &&
+        part.text &&
+        !('thought' in part && part.thought)
+      ) {
+        contentParts.push(part.text);
+      } else if (
+        'text' in part &&
+        part.text &&
+        'thought' in part &&
+        part.thought
+      ) {
+        thoughtParts.push(part.text);
+      } else if ('functionCall' in part && part.functionCall) {
+        functionCalls.push(part.functionCall);
+      } else if ('functionResponse' in part && part.functionResponse) {
+        functionResponses.push(part.functionResponse);
+      } else if ('inlineData' in part && part.inlineData) {
+        const { data, mimeType } = part.inlineData;
+        if (data && mimeType) {
+          const mediaType = this.getMediaType(mimeType);
+          mediaParts.push({ type: mediaType, data, mimeType });
+        }
+      } else if ('fileData' in part && part.fileData) {
+        const { fileUri, mimeType } = part.fileData;
+        if (fileUri && mimeType) {
+          const mediaType = this.getMediaType(mimeType);
+          mediaParts.push({
+            type: mediaType,
+            data: '',
+            mimeType,
+            fileUri,
+          });
+        }
+      }
+    }
+
+    return {
+      thoughtParts,
+      contentParts,
+      functionCalls,
+      functionResponses,
+      mediaParts,
+    };
+  }
+
  private extractFunctionResponseContent(response: unknown): string {
    if (response === null || response === undefined) {
      return '';
@@ -530,96 +535,6 @@ export class OpenAIContentConverter {
    }
  }

-  /**
-   * Create a tool message from function response (with embedded media parts)
-   */
-  private createToolMessage(
-    response: FunctionResponse,
-  ): OpenAI.Chat.ChatCompletionToolMessageParam | null {
-    const textContent = this.extractFunctionResponseContent(response.response);
-    const contentParts: OpenAIContentPart[] = [];
-
-    // Add text content first if present
-    if (textContent) {
-      contentParts.push({ type: 'text' as const, text: textContent });
-    }
-
-    // Add media parts from function response
-    for (const part of response.parts || []) {
-      const mediaPart = this.createMediaContentPart(part);
-      if (mediaPart) {
-        contentParts.push(mediaPart);
-      }
-    }
-
-    // Tool messages require content, so skip if empty
-    if (contentParts.length === 0) {
-      return null;
-    }
-
-    // Cast to OpenAI type - some OpenAI-compatible APIs support richer content in tool messages
-    return {
-      role: 'tool' as const,
-      tool_call_id: response.id || '',
-      content: contentParts as unknown as
-        | string
-        | OpenAI.Chat.ChatCompletionContentPartText[],
-    };
-  }
-
-  /**
-   * Create OpenAI media content part from Gemini part
-   */
-  private createMediaContentPart(part: Part): OpenAIContentPart | null {
-    if (part.inlineData?.mimeType && part.inlineData?.data) {
-      const mediaType = this.getMediaType(part.inlineData.mimeType);
-      if (mediaType === 'image') {
-        const dataUrl = `data:${part.inlineData.mimeType};base64,${part.inlineData.data}`;
-        return {
-          type: 'image_url' as const,
-          image_url: { url: dataUrl },
-        };
-      }
-      if (mediaType === 'audio') {
-        const format = this.getAudioFormat(part.inlineData.mimeType);
-        if (format) {
-          return {
-            type: 'input_audio' as const,
-            input_audio: {
-              data: part.inlineData.data,
-              format,
-            },
-          };
-        }
-      }
-    }
-
-    if (part.fileData?.mimeType && part.fileData?.fileUri) {
-      const filename = part.fileData.displayName || 'file';
-      const fileUri = part.fileData.fileUri;
-
-      if (fileUri.startsWith('data:')) {
-        return {
-          type: 'file' as const,
-          file: {
-            filename,
-            file_data: fileUri,
-          },
-        };
-      }
-
-      return {
-        type: 'file' as const,
-        file: {
-          filename,
-          file_data: `data:${part.fileData.mimeType};base64,${fileUri}`,
-        },
-      };
-    }
-
-    return null;
-  }
-
  /**
   * Determine media type from MIME type
   */
@@ -629,6 +544,85 @@ export class OpenAIContentConverter {
    return 'file';
  }

+  /**
+   * Create multimodal OpenAI message from parsed parts
+   */
+  private createMultimodalMessage(
+    role: 'user' | 'assistant',
+    parsedParts: Pick<
+      ParsedParts,
+      'contentParts' | 'mediaParts' | 'thoughtParts'
+    >,
+  ): ExtendedChatCompletionMessageParam | null {
+    const { contentParts, mediaParts, thoughtParts } = parsedParts;
+    const reasoningContent = thoughtParts.join('');
+    const content = contentParts.map((text) => ({
+      type: 'text' as const,
+      text,
+    }));
+
+    // If no media parts, return simple text message
+    if (mediaParts.length === 0) {
+      if (content.length === 0) return null;
+      const message: ExtendedChatCompletionMessageParam = { role, content };
+      // Only include reasoning_content if it has actual content
+      if (reasoningContent) {
+        (
+          message as ExtendedChatCompletionAssistantMessageParam
+        ).reasoning_content = reasoningContent;
+      }
+      return message;
+    }
+
+    // For assistant messages with media, convert to text only
+    // since OpenAI assistant messages don't support media content arrays
+    if (role === 'assistant') {
+      return content.length > 0
+        ? { role: 'assistant' as const, content }
+        : null;
+    }
+
+    const contentArray: OpenAI.Chat.ChatCompletionContentPart[] = [...content];
+
+    // Add media content
+    for (const mediaPart of mediaParts) {
+      if (mediaPart.type === 'image') {
+        if (mediaPart.fileUri) {
+          // For file URIs, use the URI directly
+          contentArray.push({
+            type: 'image_url' as const,
+            image_url: { url: mediaPart.fileUri },
+          });
+        } else if (mediaPart.data) {
+          // For inline data, create data URL
+          const dataUrl = `data:${mediaPart.mimeType};base64,${mediaPart.data}`;
+          contentArray.push({
+            type: 'image_url' as const,
+            image_url: { url: dataUrl },
+          });
+        }
+      } else if (mediaPart.type === 'audio' && mediaPart.data) {
+        // Convert audio format from MIME type
+        const format = this.getAudioFormat(mediaPart.mimeType);
+        if (format) {
+          contentArray.push({
+            type: 'input_audio' as const,
+            input_audio: {
+              data: mediaPart.data,
+              format: format as 'wav' | 'mp3',
+            },
+          });
+        }
+      }
+      // Note: File type is not directly supported in OpenAI's current API
+      // Could be extended in the future or handled as text description
+    }
+
+    return contentArray.length > 0
+      ? { role: 'user' as const, content: contentArray }
+      : null;
+  }
+
  /**
   * Convert MIME type to OpenAI audio format
   */
@@ -699,9 +693,8 @@ export class OpenAIContentConverter {
    const parts: Part[] = [];

    // Handle reasoning content (thoughts)
-    const reasoningText =
-      (choice.message as ExtendedCompletionMessage).reasoning_content ??
-      (choice.message as ExtendedCompletionMessage).reasoning;
+    const reasoningText = (choice.message as ExtendedCompletionMessage)
+      .reasoning_content;
    if (reasoningText) {
      parts.push({ text: reasoningText, thought: true });
    }
@@ -805,9 +798,8 @@ export class OpenAIContentConverter {
    if (choice) {
      const parts: Part[] = [];

-      const reasoningText =
-        (choice.delta as ExtendedCompletionChunkDelta)?.reasoning_content ??
-        (choice.delta as ExtendedCompletionChunkDelta)?.reasoning;
+      const reasoningText = (choice.delta as ExtendedCompletionChunkDelta)
+        ?.reasoning_content;
      if (reasoningText) {
        parts.push({ text: reasoningText, thought: true });
      }
@@ -1138,10 +1130,6 @@ export class OpenAIContentConverter {

        // If the last message is also an assistant message, merge them
        if (lastMessage.role === 'assistant') {
-          const lastToolCalls =
-            'tool_calls' in lastMessage ? lastMessage.tool_calls || [] : [];
-          const currentToolCalls =
-            'tool_calls' in message ? message.tool_calls || [] : [];
          // Combine content
          const lastContent = lastMessage.content;
          const currentContent = message.content;
@@ -1183,6 +1171,10 @@ export class OpenAIContentConverter {
          }

          // Combine tool calls
+          const lastToolCalls =
+            'tool_calls' in lastMessage ? lastMessage.tool_calls || [] : [];
+          const currentToolCalls =
+            'tool_calls' in message ? message.tool_calls || [] : [];
          const combinedToolCalls = [...lastToolCalls, ...currentToolCalls];

          // Update the last message with combined data
--- a/packages/core/src/core/openaiContentGenerator/pipeline.ts
+++ b/packages/core/src/core/openaiContentGenerator/pipeline.ts
@@ -320,15 +320,13 @@ export class ContentGenerationPipeline {
        'frequency_penalty',
        'frequencyPenalty',
      ),
-      ...this.buildReasoningConfig(request),
+      ...this.buildReasoningConfig(),
    };

    return params;
  }

-  private buildReasoningConfig(
-    request: GenerateContentParameters,
-  ): Record<string, unknown> {
+  private buildReasoningConfig(): Record<string, unknown> {
    // Reasoning configuration for OpenAI-compatible endpoints is highly fragmented.
    // For example, across common providers and models:
    //
@@ -338,21 +336,13 @@ export class ContentGenerationPipeline {
    //   - gpt-5.x series      — thinking is enabled by default; can be disabled via `reasoning.effort`
    //   - qwen3 series        — model-dependent; can be manually disabled via `extra_body.enable_thinking`
    //
-    // Given this inconsistency, we avoid mapping values and only pass through the
-    // configured reasoning object when explicitly enabled. This keeps provider- and
-    // model-specific semantics intact while honoring request-level opt-out.
+    // Given this inconsistency, we choose not to set any reasoning config here and
+    // instead rely on each model’s default behavior.

-    if (request.config?.thinkingConfig?.includeThoughts === false) {
-      return {};
-    }
+    // We plan to introduce provider- and model-specific settings to enable more
+    // fine-grained control over reasoning configuration.

-    const reasoning = this.contentGeneratorConfig.reasoning;
-
-    if (reasoning === false || reasoning === undefined) {
-      return {};
-    }
-
-    return { reasoning };
+    return {};
  }

  /**
--- a/packages/core/src/core/openaiContentGenerator/provider/dashscope.test.ts
+++ b/packages/core/src/core/openaiContentGenerator/provider/dashscope.test.ts
@@ -608,7 +608,7 @@ describe('DashScopeOpenAICompatibleProvider', () => {
      });
    });

-    it('should add cache control to last item even if not text for streaming requests', () => {
+    it('should add empty text item with cache control if last item is not text for streaming requests', () => {
      const requestWithNonTextLast: OpenAI.Chat.ChatCompletionCreateParams = {
        model: 'qwen-max',
        stream: true, // This will trigger cache control on last message
@@ -633,12 +633,12 @@ describe('DashScopeOpenAICompatibleProvider', () => {

      const content = result.messages[0]
        .content as OpenAI.Chat.ChatCompletionContentPart[];
-      expect(content).toHaveLength(2);
+      expect(content).toHaveLength(3);

-      // Cache control should be added to the last item (image)
-      expect(content[1]).toEqual({
-        type: 'image_url',
-        image_url: { url: 'https://example.com/image.jpg' },
+      // Should add empty text item with cache control
+      expect(content[2]).toEqual({
+        type: 'text',
+        text: '',
        cache_control: { type: 'ephemeral' },
      });
    });
@@ -709,8 +709,13 @@ describe('DashScopeOpenAICompatibleProvider', () => {

      const content = result.messages[0]
        .content as OpenAI.Chat.ChatCompletionContentPart[];
-      // Empty content array should remain empty
-      expect(content).toEqual([]);
+      expect(content).toEqual([
+        {
+          type: 'text',
+          text: '',
+          cache_control: { type: 'ephemeral' },
+        },
+      ]);
    });
  });

--- a/packages/core/src/core/openaiContentGenerator/provider/dashscope.ts
+++ b/packages/core/src/core/openaiContentGenerator/provider/dashscope.ts
@@ -257,15 +257,31 @@ export class DashScopeOpenAICompatibleProvider
    contentArray: ChatCompletionContentPartWithCache[],
  ): ChatCompletionContentPartWithCache[] {
    if (contentArray.length === 0) {
-      return contentArray;
+      return [
+        {
+          type: 'text',
+          text: '',
+          cache_control: { type: 'ephemeral' },
+        } as ChatCompletionContentPartTextWithCache,
+      ];
    }

-    // Add cache_control to the last text item
    const lastItem = contentArray[contentArray.length - 1];
-    contentArray[contentArray.length - 1] = {
-      ...lastItem,
-      cache_control: { type: 'ephemeral' },
-    } as ChatCompletionContentPartTextWithCache;
+
+    if (lastItem.type === 'text') {
+      // Add cache_control to the last text item
+      contentArray[contentArray.length - 1] = {
+        ...lastItem,
+        cache_control: { type: 'ephemeral' },
+      } as ChatCompletionContentPartTextWithCache;
+    } else {
+      // If the last item is not text, add a new text item with cache_control
+      contentArray.push({
+        type: 'text',
+        text: '',
+        cache_control: { type: 'ephemeral' },
+      } as ChatCompletionContentPartTextWithCache);
+    }

    return contentArray;
  }
--- a/packages/core/src/models/constants.ts
+++ b/packages/core/src/models/constants.ts
@@ -102,14 +102,16 @@ export const QWEN_OAUTH_ALLOWED_MODELS = [
 export const QWEN_OAUTH_MODELS: ModelConfig[] = [
  {
    id: 'coder-model',
-    name: 'coder-model',
-    description: 'The latest Qwen Coder model from Alibaba Cloud ModelStudio',
+    name: 'Qwen Coder',
+    description:
+      'The latest Qwen Coder model from Alibaba Cloud ModelStudio (version: qwen3-coder-plus-2025-09-23)',
    capabilities: { vision: false },
  },
  {
    id: 'vision-model',
-    name: 'vision-model',
-    description: 'The latest Qwen Vision model from Alibaba Cloud ModelStudio',
+    name: 'Qwen Vision',
+    description:
+      'The latest Qwen Vision model from Alibaba Cloud ModelStudio (version: qwen3-vl-plus-2025-09-23)',
    capabilities: { vision: true },
  },
 ];
--- a/packages/core/src/models/modelsConfig.test.ts
+++ b/packages/core/src/models/modelsConfig.test.ts
@@ -191,7 +191,7 @@ describe('ModelsConfig', () => {
    expect(gc.apiKeyEnvKey).toBe('API_KEY_SHARED');
  });

-  it('should use provider config when modelId exists in registry even after updateCredentials', () => {
+  it('should preserve settings generationConfig when model is updated via updateCredentials even if it matches modelProviders', () => {
    const modelProvidersConfig: ModelProvidersConfig = {
      openai: [
        {
@@ -213,7 +213,7 @@ describe('ModelsConfig', () => {
      initialAuthType: AuthType.USE_OPENAI,
      modelProvidersConfig,
      generationConfig: {
-        model: 'custom-model',
+        model: 'model-a',
        samplingParams: { temperature: 0.9, max_tokens: 999 },
        timeout: 9999,
        maxRetries: 9,
@@ -235,30 +235,30 @@ describe('ModelsConfig', () => {
      },
    });

-    // User manually updates credentials via updateCredentials.
-    // Note: In practice, handleAuthSelect prevents using a modelId that matches a provider model,
-    // but if syncAfterAuthRefresh is called with a modelId that exists in registry,
-    // we should use provider config.
-    modelsConfig.updateCredentials({ apiKey: 'manual-key' });
+    // User manually updates the model via updateCredentials (e.g. key prompt flow).
+    // Even if the model ID matches a modelProviders entry, we must not apply provider defaults
+    // that would overwrite settings.model.generationConfig.
+    modelsConfig.updateCredentials({ model: 'model-a' });

-    // syncAfterAuthRefresh with a modelId that exists in registry should use provider config
-    modelsConfig.syncAfterAuthRefresh(AuthType.USE_OPENAI, 'model-a');
+    modelsConfig.syncAfterAuthRefresh(
+      AuthType.USE_OPENAI,
+      modelsConfig.getModel(),
+    );

    const gc = currentGenerationConfig(modelsConfig);
    expect(gc.model).toBe('model-a');
-    // Provider config should be applied
-    expect(gc.samplingParams?.temperature).toBe(0.1);
-    expect(gc.samplingParams?.max_tokens).toBe(123);
-    expect(gc.timeout).toBe(111);
-    expect(gc.maxRetries).toBe(1);
+    expect(gc.samplingParams?.temperature).toBe(0.9);
+    expect(gc.samplingParams?.max_tokens).toBe(999);
+    expect(gc.timeout).toBe(9999);
+    expect(gc.maxRetries).toBe(9);
  });

-  it('should preserve settings generationConfig when modelId does not exist in registry', () => {
+  it('should preserve settings generationConfig across multiple auth refreshes after updateCredentials', () => {
    const modelProvidersConfig: ModelProvidersConfig = {
      openai: [
        {
-          id: 'provider-model',
-          name: 'Provider Model',
+          id: 'model-a',
+          name: 'Model A',
          baseUrl: 'https://api.example.com/v1',
          envKey: 'API_KEY_A',
          generationConfig: {
@@ -270,12 +270,11 @@ describe('ModelsConfig', () => {
      ],
    };

-    // Simulate settings with a custom model (not in registry)
    const modelsConfig = new ModelsConfig({
      initialAuthType: AuthType.USE_OPENAI,
      modelProvidersConfig,
      generationConfig: {
-        model: 'custom-model',
+        model: 'model-a',
        samplingParams: { temperature: 0.9, max_tokens: 999 },
        timeout: 9999,
        maxRetries: 9,
@@ -297,21 +296,25 @@ describe('ModelsConfig', () => {
      },
    });

-    // User manually sets credentials for a custom model (not in registry)
    modelsConfig.updateCredentials({
      apiKey: 'manual-key',
      baseUrl: 'https://manual.example.com/v1',
-      model: 'custom-model',
+      model: 'model-a',
    });

-    // First auth refresh - modelId doesn't exist in registry, so credentials should be preserved
-    modelsConfig.syncAfterAuthRefresh(AuthType.USE_OPENAI, 'custom-model');
+    // First auth refresh
+    modelsConfig.syncAfterAuthRefresh(
+      AuthType.USE_OPENAI,
+      modelsConfig.getModel(),
+    );
    // Second auth refresh should still preserve settings generationConfig
-    modelsConfig.syncAfterAuthRefresh(AuthType.USE_OPENAI, 'custom-model');
+    modelsConfig.syncAfterAuthRefresh(
+      AuthType.USE_OPENAI,
+      modelsConfig.getModel(),
+    );

    const gc = currentGenerationConfig(modelsConfig);
-    expect(gc.model).toBe('custom-model');
-    // Settings-sourced generation config should be preserved since modelId doesn't exist in registry
+    expect(gc.model).toBe('model-a');
    expect(gc.samplingParams?.temperature).toBe(0.9);
    expect(gc.samplingParams?.max_tokens).toBe(999);
    expect(gc.timeout).toBe(9999);
@@ -678,120 +681,4 @@ describe('ModelsConfig', () => {
    expect(modelsConfig.getModel()).toBe('updated-model');
    expect(modelsConfig.getGenerationConfig().model).toBe('updated-model');
  });
-
-  describe('getAllAvailableModels', () => {
-    it('should return all models across all authTypes', () => {
-      const modelProvidersConfig: ModelProvidersConfig = {
-        openai: [
-          {
-            id: 'openai-model-1',
-            name: 'OpenAI Model 1',
-            baseUrl: 'https://api.openai.com/v1',
-            envKey: 'OPENAI_API_KEY',
-          },
-          {
-            id: 'openai-model-2',
-            name: 'OpenAI Model 2',
-            baseUrl: 'https://api.openai.com/v1',
-            envKey: 'OPENAI_API_KEY',
-          },
-        ],
-        anthropic: [
-          {
-            id: 'anthropic-model-1',
-            name: 'Anthropic Model 1',
-            baseUrl: 'https://api.anthropic.com/v1',
-            envKey: 'ANTHROPIC_API_KEY',
-          },
-        ],
-        gemini: [
-          {
-            id: 'gemini-model-1',
-            name: 'Gemini Model 1',
-            baseUrl: 'https://generativelanguage.googleapis.com/v1',
-            envKey: 'GEMINI_API_KEY',
-          },
-        ],
-      };
-
-      const modelsConfig = new ModelsConfig({
-        modelProvidersConfig,
-      });
-
-      const allModels = modelsConfig.getAllAvailableModels();
-
-      // Should include qwen-oauth models (hard-coded)
-      const qwenModels = allModels.filter(
-        (m) => m.authType === AuthType.QWEN_OAUTH,
-      );
-      expect(qwenModels.length).toBeGreaterThan(0);
-
-      // Should include openai models
-      const openaiModels = allModels.filter(
-        (m) => m.authType === AuthType.USE_OPENAI,
-      );
-      expect(openaiModels.length).toBe(2);
-      expect(openaiModels.map((m) => m.id)).toContain('openai-model-1');
-      expect(openaiModels.map((m) => m.id)).toContain('openai-model-2');
-
-      // Should include anthropic models
-      const anthropicModels = allModels.filter(
-        (m) => m.authType === AuthType.USE_ANTHROPIC,
-      );
-      expect(anthropicModels.length).toBe(1);
-      expect(anthropicModels[0].id).toBe('anthropic-model-1');
-
-      // Should include gemini models
-      const geminiModels = allModels.filter(
-        (m) => m.authType === AuthType.USE_GEMINI,
-      );
-      expect(geminiModels.length).toBe(1);
-      expect(geminiModels[0].id).toBe('gemini-model-1');
-    });
-
-    it('should return empty array when no models are registered', () => {
-      const modelsConfig = new ModelsConfig();
-
-      const allModels = modelsConfig.getAllAvailableModels();
-
-      // Should still include qwen-oauth models (hard-coded)
-      expect(allModels.length).toBeGreaterThan(0);
-      const qwenModels = allModels.filter(
-        (m) => m.authType === AuthType.QWEN_OAUTH,
-      );
-      expect(qwenModels.length).toBeGreaterThan(0);
-    });
-
-    it('should return models with correct structure', () => {
-      const modelProvidersConfig: ModelProvidersConfig = {
-        openai: [
-          {
-            id: 'test-model',
-            name: 'Test Model',
-            description: 'A test model',
-            baseUrl: 'https://api.example.com/v1',
-            envKey: 'TEST_API_KEY',
-            capabilities: {
-              vision: true,
-            },
-          },
-        ],
-      };
-
-      const modelsConfig = new ModelsConfig({
-        modelProvidersConfig,
-      });
-
-      const allModels = modelsConfig.getAllAvailableModels();
-      const testModel = allModels.find((m) => m.id === 'test-model');
-
-      expect(testModel).toBeDefined();
-      expect(testModel?.id).toBe('test-model');
-      expect(testModel?.label).toBe('Test Model');
-      expect(testModel?.description).toBe('A test model');
-      expect(testModel?.authType).toBe(AuthType.USE_OPENAI);
-      expect(testModel?.isVision).toBe(true);
-      expect(testModel?.capabilities?.vision).toBe(true);
-    });
-  });
 });
--- a/packages/core/src/models/modelsConfig.ts
+++ b/packages/core/src/models/modelsConfig.ts
@@ -203,18 +203,6 @@ export class ModelsConfig {
    return this.modelRegistry.getModelsForAuthType(authType);
  }

-  /**
-   * Get all available models across all authTypes
-   */
-  getAllAvailableModels(): AvailableModel[] {
-    const allModels: AvailableModel[] = [];
-    for (const authType of Object.values(AuthType)) {
-      const models = this.modelRegistry.getModelsForAuthType(authType);
-      allModels.push(...models);
-    }
-    return allModels;
-  }
-
  /**
   * Check if a model exists for the given authType
   */
@@ -319,33 +307,6 @@ export class ModelsConfig {
    return this.generationConfigSources;
  }

-  /**
-   * Merge settings generation config, preserving existing values.
-   * Used when provider-sourced config is cleared but settings should still apply.
-   */
-  mergeSettingsGenerationConfig(
-    settingsGenerationConfig?: Partial<ContentGeneratorConfig>,
-  ): void {
-    if (!settingsGenerationConfig) {
-      return;
-    }
-
-    for (const field of MODEL_GENERATION_CONFIG_FIELDS) {
-      if (
-        !(field in this._generationConfig) &&
-        field in settingsGenerationConfig
-      ) {
-        // eslint-disable-next-line @typescript-eslint/no-explicit-any
-        (this._generationConfig as any)[field] =
-          settingsGenerationConfig[field];
-        this.generationConfigSources[field] = {
-          kind: 'settings',
-          detail: `model.generationConfig.${field}`,
-        };
-      }
-    }
-  }
-
  /**
   * Update credentials in generation config.
   * Sets a flag to prevent syncAfterAuthRefresh from overriding these credentials.
@@ -353,20 +314,12 @@ export class ModelsConfig {
   * When credentials are manually set, we clear all provider-sourced configuration
   * to maintain provider atomicity (either fully applied or not at all).
   * Other layers (CLI, env, settings, defaults) will participate in resolve.
-   *
-   * @param settingsGenerationConfig Optional generation config from settings.json
-   *                                  to merge after clearing provider-sourced config.
-   *                                  This ensures settings.model.generationConfig fields
-   *                                  (e.g., samplingParams, timeout) are preserved.
   */
-  updateCredentials(
-    credentials: {
-      apiKey?: string;
-      baseUrl?: string;
-      model?: string;
-    },
-    settingsGenerationConfig?: Partial<ContentGeneratorConfig>,
-  ): void {
+  updateCredentials(credentials: {
+    apiKey?: string;
+    baseUrl?: string;
+    model?: string;
+  }): void {
    /**
     * If any fields are updated here, we treat the resulting config as manually overridden
     * and avoid applying modelProvider defaults during the next auth refresh.
@@ -406,14 +359,6 @@ export class ModelsConfig {
    this.strictModelProviderSelection = false;
    // Clear apiKeyEnvKey to prevent validation from requiring environment variable
    this._generationConfig.apiKeyEnvKey = undefined;
-
-    // After clearing provider-sourced config, merge settings.model.generationConfig
-    // to ensure fields like samplingParams, timeout, etc. are preserved.
-    // This follows the resolution strategy where settings.model.generationConfig
-    // has lower priority than programmatic overrides but should still be applied.
-    if (settingsGenerationConfig) {
-      this.mergeSettingsGenerationConfig(settingsGenerationConfig);
-    }
  }

  /**
@@ -642,88 +587,50 @@ export class ModelsConfig {
  }

  /**
-   * Sync state after auth refresh with fallback strategy:
-   * 1. If modelId can be found in modelRegistry, use the config from modelRegistry.
-   * 2. Otherwise, if existing credentials exist in resolved generationConfig from other sources
-   *    (not modelProviders), preserve them and update authType/modelId only.
-   * 3. Otherwise, fall back to default model for the authType.
-   * 4. If no default is available, leave the generationConfig incomplete and let
-   *    resolveContentGeneratorConfigWithSources throw exceptions as expected.
+   * Called by Config.refreshAuth to sync state after auth refresh.
+   *
+   * IMPORTANT: If credentials were manually set via updateCredentials(),
+   * we should NOT override them with modelProvider defaults.
+   * This handles the case where user inputs credentials via OpenAIKeyPrompt
+   * after removing environment variables for a previously selected model.
   */
  syncAfterAuthRefresh(authType: AuthType, modelId?: string): void {
-    this.strictModelProviderSelection = false;
-    const previousAuthType = this.currentAuthType;
-    this.currentAuthType = authType;
+    // Check if we have manually set credentials that should be preserved
+    const preserveManualCredentials = this.hasManualCredentials;
+
+    // If credentials were manually set, don't apply modelProvider defaults
+    // Just update the authType and preserve the manually set credentials
+    if (preserveManualCredentials && authType === AuthType.USE_OPENAI) {
+      this.strictModelProviderSelection = false;
+      this.currentAuthType = authType;
+      if (modelId) {
+        this._generationConfig.model = modelId;
+      }
+      return;
+    }
+
+    this.strictModelProviderSelection = false;

-    // Step 1: If modelId exists in registry, always use config from modelRegistry
-    // Manual credentials won't have a modelId that matches a provider model (handleAuthSelect prevents it),
-    // so if modelId exists in registry, we should always use provider config.
-    // This handles provider switching even within the same authType.
    if (modelId && this.modelRegistry.hasModel(authType, modelId)) {
      const resolved = this.modelRegistry.getModel(authType, modelId);
      if (resolved) {
+        // Ensure applyResolvedModelDefaults can correctly apply authType-specific
+        // behavior (e.g., Qwen OAuth placeholder token) by setting currentAuthType
+        // before applying defaults.
+        this.currentAuthType = authType;
        this.applyResolvedModelDefaults(resolved);
-        this.strictModelProviderSelection = true;
-        return;
      }
-    }
-
-    // Step 2: Check if there are existing credentials from other sources (not modelProviders)
-    const apiKeySource = this.generationConfigSources['apiKey'];
-    const baseUrlSource = this.generationConfigSources['baseUrl'];
-    const hasExistingCredentials =
-      (this._generationConfig.apiKey &&
-        apiKeySource?.kind !== 'modelProviders') ||
-      (this._generationConfig.baseUrl &&
-        baseUrlSource?.kind !== 'modelProviders');
-
-    // Only preserve credentials if:
-    // 1. AuthType hasn't changed (credentials are authType-specific), AND
-    // 2. The modelId doesn't exist in the registry (if it did, we would have used provider config in Step 1), AND
-    // 3. Either:
-    //    a. We have manual credentials (set via updateCredentials), OR
-    //    b. We have existing credentials
-    // Note: Even if authType hasn't changed, switching to a different provider model (that exists in registry)
-    // will use provider config (Step 1), not preserve old credentials. This ensures credentials change when
-    // switching providers, independent of authType changes.
-    const isAuthTypeChange = previousAuthType !== authType;
-    const shouldPreserveCredentials =
-      !isAuthTypeChange &&
-      (modelId === undefined ||
-        !this.modelRegistry.hasModel(authType, modelId)) &&
-      (this.hasManualCredentials || hasExistingCredentials);
-
-    if (shouldPreserveCredentials) {
-      // Preserve existing credentials, just update authType and modelId if provided
-      if (modelId) {
-        this._generationConfig.model = modelId;
-        if (!this.generationConfigSources['model']) {
-          this.generationConfigSources['model'] = {
-            kind: 'programmatic',
-            detail: 'auth refresh (preserved credentials)',
-          };
-        }
-      }
-      return;
-    }
-
-    // Step 3: Fall back to default model for the authType
-    const defaultModel =
-      this.modelRegistry.getDefaultModelForAuthType(authType);
-    if (defaultModel) {
-      this.applyResolvedModelDefaults(defaultModel);
-      return;
-    }
-
-    // Step 4: No default available - leave generationConfig incomplete
-    // resolveContentGeneratorConfigWithSources will throw exceptions as expected
-    if (modelId) {
-      this._generationConfig.model = modelId;
-      if (!this.generationConfigSources['model']) {
-        this.generationConfigSources['model'] = {
-          kind: 'programmatic',
-          detail: 'auth refresh (no default model)',
-        };
+    } else {
+      // If the provided modelId doesn't exist in the registry for the new authType,
+      // use the default model for that authType instead of keeping the old model.
+      // This handles the case where switching from one authType (e.g., OPENAI with
+      // env vars) to another (e.g., qwen-oauth) - we should use the default model
+      // for the new authType, not the old model.
+      this.currentAuthType = authType;
+      const defaultModel =
+        this.modelRegistry.getDefaultModelForAuthType(authType);
+      if (defaultModel) {
+        this.applyResolvedModelDefaults(defaultModel);
      }
    }
  }
--- a/packages/core/src/qwen/qwenOAuth2.test.ts
+++ b/packages/core/src/qwen/qwenOAuth2.test.ts
@@ -751,7 +751,6 @@ describe('getQwenOAuthClient', () => {
  beforeEach(() => {
    mockConfig = {
      isBrowserLaunchSuppressed: vi.fn().mockReturnValue(false),
-      isInteractive: vi.fn().mockReturnValue(true),
    } as unknown as Config;

    originalFetch = global.fetch;
@@ -840,7 +839,9 @@ describe('getQwenOAuthClient', () => {
          requireCachedCredentials: true,
        }),
      ),
-    ).rejects.toThrow('Please use /auth to re-authenticate.');
+    ).rejects.toThrow(
+      'No cached Qwen-OAuth credentials found. Please re-authenticate.',
+    );

    expect(global.fetch).not.toHaveBeenCalled();

@@ -1006,7 +1007,6 @@ describe('getQwenOAuthClient - Enhanced Error Scenarios', () => {
  beforeEach(() => {
    mockConfig = {
      isBrowserLaunchSuppressed: vi.fn().mockReturnValue(false),
-      isInteractive: vi.fn().mockReturnValue(true),
    } as unknown as Config;

    originalFetch = global.fetch;
@@ -1202,7 +1202,6 @@ describe('authWithQwenDeviceFlow - Comprehensive Testing', () => {
  beforeEach(() => {
    mockConfig = {
      isBrowserLaunchSuppressed: vi.fn().mockReturnValue(false),
-      isInteractive: vi.fn().mockReturnValue(true),
    } as unknown as Config;

    originalFetch = global.fetch;
@@ -1406,7 +1405,6 @@ describe('Browser Launch and Error Handling', () => {
  beforeEach(() => {
    mockConfig = {
      isBrowserLaunchSuppressed: vi.fn().mockReturnValue(false),
-      isInteractive: vi.fn().mockReturnValue(true),
    } as unknown as Config;

    originalFetch = global.fetch;
@@ -2045,7 +2043,6 @@ describe('SharedTokenManager Integration in QwenOAuth2Client', () => {
  it('should handle TokenManagerError types correctly in getQwenOAuthClient', async () => {
    const mockConfig = {
      isBrowserLaunchSuppressed: vi.fn().mockReturnValue(true),
-      isInteractive: vi.fn().mockReturnValue(true),
    } as unknown as Config;

    // Test different TokenManagerError types
--- a/packages/core/src/qwen/qwenOAuth2.ts
+++ b/packages/core/src/qwen/qwenOAuth2.ts
@@ -516,7 +516,9 @@ export async function getQwenOAuthClient(
    }

    if (options?.requireCachedCredentials) {
-      throw new Error('Please use /auth to re-authenticate.');
+      throw new Error(
+        'No cached Qwen-OAuth credentials found. Please re-authenticate.',
+      );
    }

    // If we couldn't obtain valid credentials via SharedTokenManager, fall back to
@@ -738,9 +740,11 @@ async function authWithQwenDeviceFlow(
    // Emit device authorization event for UI integration immediately
    qwenOAuth2Events.emit(QwenOAuth2Event.AuthUri, deviceAuth);

-    if (config.isBrowserLaunchSuppressed() || !config.isInteractive()) {
-      showFallbackMessage(deviceAuth.verification_uri_complete);
-    }
+    // Always show the fallback message in non-interactive environments to ensure
+    // users can see the authorization URL even if browser launching is attempted.
+    // This is critical for headless/remote environments where browser launching
+    // may silently fail without throwing an error.
+    showFallbackMessage(deviceAuth.verification_uri_complete);

    // Try to open browser if not suppressed
    if (!config.isBrowserLaunchSuppressed()) {
--- a/packages/core/src/skills/skill-manager.test.ts
+++ b/packages/core/src/skills/skill-manager.test.ts
@@ -112,62 +112,6 @@ You are a helpful assistant with this skill.
      expect(config.filePath).toBe(validSkillConfig.filePath);
    });

-    it('should parse markdown with CRLF line endings', () => {
-      const markdownCrlf = `---\r
-name: test-skill\r
-description: A test skill\r
---\r
-\r
-You are a helpful assistant with this skill.\r
-`;
-
-      const config = manager.parseSkillContent(
-        markdownCrlf,
-        validSkillConfig.filePath,
-        'project',
-      );
-
-      expect(config.name).toBe('test-skill');
-      expect(config.description).toBe('A test skill');
-      expect(config.body).toBe('You are a helpful assistant with this skill.');
-    });
-
-    it('should parse markdown with UTF-8 BOM', () => {
-      const markdownWithBom = `\uFEFF---
-name: test-skill
-description: A test skill
---
-
-You are a helpful assistant with this skill.
-`;
-
-      const config = manager.parseSkillContent(
-        markdownWithBom,
-        validSkillConfig.filePath,
-        'project',
-      );
-
-      expect(config.name).toBe('test-skill');
-      expect(config.description).toBe('A test skill');
-    });
-
-    it('should parse markdown when body is empty and file ends after frontmatter', () => {
-      const frontmatterOnly = `---
-name: test-skill
-description: A test skill
---`;
-
-      const config = manager.parseSkillContent(
-        frontmatterOnly,
-        validSkillConfig.filePath,
-        'project',
-      );
-
-      expect(config.name).toBe('test-skill');
-      expect(config.description).toBe('A test skill');
-      expect(config.body).toBe('');
-    });
-
    it('should parse content with allowedTools', () => {
      const markdownWithTools = `---
 name: test-skill
--- a/packages/core/src/skills/skill-manager.ts
+++ b/packages/core/src/skills/skill-manager.ts
@@ -235,7 +235,6 @@ export class SkillManager {
    }

    this.watchStarted = true;
-    await this.ensureUserSkillsDir();
    await this.refreshCache();
    this.updateWatchersFromCache();
  }
@@ -307,11 +306,9 @@ export class SkillManager {
    level: SkillLevel,
  ): SkillConfig {
    try {
-      const normalizedContent = normalizeSkillFileContent(content);
-
      // Split frontmatter and content
-      const frontmatterRegex = /^---\n([\s\S]*?)\n---(?:\n|$)([\s\S]*)$/;
-      const match = normalizedContent.match(frontmatterRegex);
+      const frontmatterRegex = /^---\n([\s\S]*?)\n---\n([\s\S]*)$/;
+      const match = content.match(frontmatterRegex);

      if (!match) {
        throw new Error('Invalid format: missing YAML frontmatter');
@@ -489,14 +486,29 @@ export class SkillManager {
  }

  private updateWatchersFromCache(): void {
-    const watchTargets = new Set<string>(
-      (['project', 'user'] as const)
-        .map((level) => this.getSkillsBaseDir(level))
-        .filter((baseDir) => fsSync.existsSync(baseDir)),
-    );
+    const desiredPaths = new Set<string>();
+
+    for (const level of ['project', 'user'] as const) {
+      const baseDir = this.getSkillsBaseDir(level);
+      const parentDir = path.dirname(baseDir);
+      if (fsSync.existsSync(parentDir)) {
+        desiredPaths.add(parentDir);
+      }
+      if (fsSync.existsSync(baseDir)) {
+        desiredPaths.add(baseDir);
+      }
+
+      const levelSkills = this.skillsCache?.get(level) || [];
+      for (const skill of levelSkills) {
+        const skillDir = path.dirname(skill.filePath);
+        if (fsSync.existsSync(skillDir)) {
+          desiredPaths.add(skillDir);
+        }
+      }
+    }

    for (const existingPath of this.watchers.keys()) {
-      if (!watchTargets.has(existingPath)) {
+      if (!desiredPaths.has(existingPath)) {
        void this.watchers
          .get(existingPath)
          ?.close()
@@ -510,7 +522,7 @@ export class SkillManager {
      }
    }

-    for (const watchPath of watchTargets) {
+    for (const watchPath of desiredPaths) {
      if (this.watchers.has(watchPath)) {
        continue;
      }
@@ -545,26 +557,4 @@ export class SkillManager {
      void this.refreshCache().then(() => this.updateWatchersFromCache());
    }, 150);
  }
-
-  private async ensureUserSkillsDir(): Promise<void> {
-    const baseDir = this.getSkillsBaseDir('user');
-    try {
-      await fs.mkdir(baseDir, { recursive: true });
-    } catch (error) {
-      console.warn(
-        `Failed to create user skills directory at ${baseDir}:`,
-        error,
-      );
-    }
-  }
-}
-
-function normalizeSkillFileContent(content: string): string {
-  // Strip UTF-8 BOM to ensure frontmatter starts at the first character.
-  let normalized = content.replace(/^\uFEFF/, '');
-
-  // Normalize line endings so skills authored on Windows (CRLF) parse correctly.
-  normalized = normalized.replace(/\r\n/g, '\n').replace(/\r/g, '\n');
-
-  return normalized;
 }
--- a/packages/core/src/telemetry/constants.ts
+++ b/packages/core/src/telemetry/constants.ts
@@ -35,7 +35,6 @@ export const EVENT_MODEL_SLASH_COMMAND = 'qwen-code.slash_command.model';
 export const EVENT_SUBAGENT_EXECUTION = 'qwen-code.subagent_execution';
 export const EVENT_SKILL_LAUNCH = 'qwen-code.skill_launch';
 export const EVENT_AUTH = 'qwen-code.auth';
-export const EVENT_USER_FEEDBACK = 'qwen-code.user_feedback';

 // Performance Events
 export const EVENT_STARTUP_PERFORMANCE = 'qwen-code.startup.performance';
--- a/packages/core/src/telemetry/index.ts
+++ b/packages/core/src/telemetry/index.ts
@@ -45,7 +45,6 @@ export {
  logNextSpeakerCheck,
  logAuth,
  logSkillLaunch,
-  logUserFeedback,
 } from './loggers.js';
 export type { SlashCommandEvent, ChatCompressionEvent } from './types.js';
 export {
@@ -66,8 +65,6 @@ export {
  NextSpeakerCheckEvent,
  AuthEvent,
  SkillLaunchEvent,
-  UserFeedbackEvent,
-  UserFeedbackRating,
 } from './types.js';
 export { makeSlashCommandEvent, makeChatCompressionEvent } from './types.js';
 export type { TelemetryEvent } from './types.js';
--- a/packages/core/src/telemetry/loggers.ts
+++ b/packages/core/src/telemetry/loggers.ts
@@ -38,7 +38,6 @@ import {
  EVENT_INVALID_CHUNK,
  EVENT_AUTH,
  EVENT_SKILL_LAUNCH,
-  EVENT_USER_FEEDBACK,
 } from './constants.js';
 import {
  recordApiErrorMetrics,
@@ -87,7 +86,6 @@ import type {
  InvalidChunkEvent,
  AuthEvent,
  SkillLaunchEvent,
-  UserFeedbackEvent,
 } from './types.js';
 import type { UiEvent } from './uiTelemetry.js';
 import { uiTelemetryService } from './uiTelemetry.js';
@@ -889,32 +887,3 @@ export function logSkillLaunch(config: Config, event: SkillLaunchEvent): void {
  };
  logger.emit(logRecord);
 }
-
-export function logUserFeedback(
-  config: Config,
-  event: UserFeedbackEvent,
-): void {
-  const uiEvent = {
-    ...event,
-    'event.name': EVENT_USER_FEEDBACK,
-    'event.timestamp': new Date().toISOString(),
-  } as UiEvent;
-  uiTelemetryService.addEvent(uiEvent);
-  config.getChatRecordingService()?.recordUiTelemetryEvent(uiEvent);
-  QwenLogger.getInstance(config)?.logUserFeedbackEvent(event);
-  if (!isTelemetrySdkInitialized()) return;
-
-  const attributes: LogAttributes = {
-    ...getCommonAttributes(config),
-    ...event,
-    'event.name': EVENT_USER_FEEDBACK,
-    'event.timestamp': new Date().toISOString(),
-  };
-
-  const logger = logs.getLogger(SERVICE_NAME);
-  const logRecord: LogRecord = {
-    body: `User feedback: Rating ${event.rating} for session ${event.session_id}.`,
-    attributes,
-  };
-  logger.emit(logRecord);
-}
--- a/packages/core/src/telemetry/qwen-logger/qwen-logger.ts
+++ b/packages/core/src/telemetry/qwen-logger/qwen-logger.ts
@@ -39,7 +39,6 @@ import type {
  ExtensionDisableEvent,
  AuthEvent,
  SkillLaunchEvent,
-  UserFeedbackEvent,
  RipgrepFallbackEvent,
  EndSessionEvent,
 } from '../types.js';
@@ -843,21 +842,6 @@ export class QwenLogger {
    this.flushIfNeeded();
  }

-  logUserFeedbackEvent(event: UserFeedbackEvent): void {
-    const rumEvent = this.createActionEvent('user', 'user_feedback', {
-      properties: {
-        session_id: event.session_id,
-        rating: event.rating,
-        model: event.model,
-        approval_mode: event.approval_mode,
-        prompt_id: event.prompt_id || '',
-      },
-    });
-
-    this.enqueueLogEvent(rumEvent);
-    this.flushIfNeeded();
-  }
-
  logChatCompressionEvent(event: ChatCompressionEvent): void {
    const rumEvent = this.createActionEvent('misc', 'chat_compression', {
      properties: {
--- a/packages/core/src/telemetry/types.ts
+++ b/packages/core/src/telemetry/types.ts
@@ -757,38 +757,6 @@ export class SkillLaunchEvent implements BaseTelemetryEvent {
  }
 }

-export enum UserFeedbackRating {
-  BAD = 1,
-  FINE = 2,
-  GOOD = 3,
-}
-
-export class UserFeedbackEvent implements BaseTelemetryEvent {
-  'event.name': 'user_feedback';
-  'event.timestamp': string;
-  session_id: string;
-  rating: UserFeedbackRating;
-  model: string;
-  approval_mode: string;
-  prompt_id?: string;
-
-  constructor(
-    session_id: string,
-    rating: UserFeedbackRating,
-    model: string,
-    approval_mode: string,
-    prompt_id?: string,
-  ) {
-    this['event.name'] = 'user_feedback';
-    this['event.timestamp'] = new Date().toISOString();
-    this.session_id = session_id;
-    this.rating = rating;
-    this.model = model;
-    this.approval_mode = approval_mode;
-    this.prompt_id = prompt_id;
-  }
-}
-
 export type TelemetryEvent =
  | StartSessionEvent
  | EndSessionEvent
@@ -818,8 +786,7 @@ export type TelemetryEvent =
  | ToolOutputTruncatedEvent
  | ModelSlashCommandEvent
  | AuthEvent
-  | SkillLaunchEvent
-  | UserFeedbackEvent;
+  | SkillLaunchEvent;

 export class ExtensionDisableEvent implements BaseTelemetryEvent {
  'event.name': 'extension_disable';
--- a/packages/core/src/tools/read-file.test.ts
+++ b/packages/core/src/tools/read-file.test.ts
@@ -283,7 +283,6 @@ describe('ReadFileTool', () => {
        inlineData: {
          data: pngHeader.toString('base64'),
          mimeType: 'image/png',
-          displayName: 'image.png',
        },
      });
      expect(result.returnDisplay).toBe('Read image file: image.png');
@@ -302,10 +301,9 @@ describe('ReadFileTool', () => {

      const result = await invocation.execute(abortSignal);
      expect(result.llmContent).toEqual({
-        fileData: {
-          fileUri: pdfHeader.toString('base64'),
+        inlineData: {
+          data: pdfHeader.toString('base64'),
          mimeType: 'application/pdf',
-          displayName: 'document.pdf',
        },
      });
      expect(result.returnDisplay).toBe('Read pdf file: document.pdf');
--- a/packages/core/src/tools/read-many-files.test.ts
+++ b/packages/core/src/tools/read-many-files.test.ts
@@ -383,7 +383,6 @@ describe('ReadManyFilesTool', () => {
              0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a,
            ]).toString('base64'),
            mimeType: 'image/png',
-            displayName: 'image.png',
          },
        },
        '\n--- End of content ---',
@@ -408,7 +407,6 @@ describe('ReadManyFilesTool', () => {
              0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a,
            ]).toString('base64'),
            mimeType: 'image/png',
-            displayName: 'myExactImage.png',
          },
        },
        '\n--- End of content ---',
@@ -436,34 +434,32 @@ describe('ReadManyFilesTool', () => {
      );
    });

-    it('should include PDF files as fileData parts if explicitly requested by extension', async () => {
+    it('should include PDF files as inlineData parts if explicitly requested by extension', async () => {
      createBinaryFile('important.pdf', Buffer.from('%PDF-1.4...'));
      const params = { paths: ['*.pdf'] }; // Explicitly requesting .pdf files
      const invocation = tool.build(params);
      const result = await invocation.execute(new AbortController().signal);
      expect(result.llmContent).toEqual([
        {
-          fileData: {
-            fileUri: Buffer.from('%PDF-1.4...').toString('base64'),
+          inlineData: {
+            data: Buffer.from('%PDF-1.4...').toString('base64'),
            mimeType: 'application/pdf',
-            displayName: 'important.pdf',
          },
        },
        '\n--- End of content ---',
      ]);
    });

-    it('should include PDF files as fileData parts if explicitly requested by name', async () => {
+    it('should include PDF files as inlineData parts if explicitly requested by name', async () => {
      createBinaryFile('report-final.pdf', Buffer.from('%PDF-1.4...'));
      const params = { paths: ['report-final.pdf'] };
      const invocation = tool.build(params);
      const result = await invocation.execute(new AbortController().signal);
      expect(result.llmContent).toEqual([
        {
-          fileData: {
-            fileUri: Buffer.from('%PDF-1.4...').toString('base64'),
+          inlineData: {
+            data: Buffer.from('%PDF-1.4...').toString('base64'),
            mimeType: 'application/pdf',
-            displayName: 'report-final.pdf',
          },
        },
        '\n--- End of content ---',
--- a/packages/core/src/tools/skill.ts
+++ b/packages/core/src/tools/skill.ts
@@ -53,7 +53,7 @@ export class SkillTool extends BaseDeclarativeTool<SkillParams, ToolResult> {
      false, // canUpdateOutput
    );

-    this.skillManager = config.getSkillManager()!;
+    this.skillManager = config.getSkillManager();
    this.skillManager.addChangeListener(() => {
      void this.refreshSkills();
    });
--- a/packages/core/src/utils/bfsFileSearch.test.ts
+++ b/packages/core/src/utils/bfsFileSearch.test.ts
@@ -0,0 +1,232 @@
+/**
+ * @license
+ * Copyright 2025 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import { describe, it, expect, beforeEach, afterEach } from 'vitest';
+import * as fsPromises from 'node:fs/promises';
+import * as path from 'node:path';
+import * as os from 'node:os';
+import { bfsFileSearch } from './bfsFileSearch.js';
+import { FileDiscoveryService } from '../services/fileDiscoveryService.js';
+
+describe('bfsFileSearch', () => {
+  let testRootDir: string;
+
+  async function createEmptyDir(...pathSegments: string[]) {
+    const fullPath = path.join(testRootDir, ...pathSegments);
+    await fsPromises.mkdir(fullPath, { recursive: true });
+    return fullPath;
+  }
+
+  async function createTestFile(content: string, ...pathSegments: string[]) {
+    const fullPath = path.join(testRootDir, ...pathSegments);
+    await fsPromises.mkdir(path.dirname(fullPath), { recursive: true });
+    await fsPromises.writeFile(fullPath, content);
+    return fullPath;
+  }
+
+  beforeEach(async () => {
+    testRootDir = await fsPromises.mkdtemp(
+      path.join(os.tmpdir(), 'bfs-file-search-test-'),
+    );
+  });
+
+  afterEach(async () => {
+    await fsPromises.rm(testRootDir, { recursive: true, force: true });
+  });
+
+  it('should find a file in the root directory', async () => {
+    const targetFilePath = await createTestFile('content', 'target.txt');
+    const result = await bfsFileSearch(testRootDir, { fileName: 'target.txt' });
+    expect(result).toEqual([targetFilePath]);
+  });
+
+  it('should find a file in a nested directory', async () => {
+    const targetFilePath = await createTestFile(
+      'content',
+      'a',
+      'b',
+      'target.txt',
+    );
+    const result = await bfsFileSearch(testRootDir, { fileName: 'target.txt' });
+    expect(result).toEqual([targetFilePath]);
+  });
+
+  it('should find multiple files with the same name', async () => {
+    const targetFilePath1 = await createTestFile('content1', 'a', 'target.txt');
+    const targetFilePath2 = await createTestFile('content2', 'b', 'target.txt');
+    const result = await bfsFileSearch(testRootDir, { fileName: 'target.txt' });
+    result.sort();
+    expect(result).toEqual([targetFilePath1, targetFilePath2].sort());
+  });
+
+  it('should return an empty array if no file is found', async () => {
+    await createTestFile('content', 'other.txt');
+    const result = await bfsFileSearch(testRootDir, { fileName: 'target.txt' });
+    expect(result).toEqual([]);
+  });
+
+  it('should ignore directories specified in ignoreDirs', async () => {
+    await createTestFile('content', 'ignored', 'target.txt');
+    const targetFilePath = await createTestFile(
+      'content',
+      'not-ignored',
+      'target.txt',
+    );
+    const result = await bfsFileSearch(testRootDir, {
+      fileName: 'target.txt',
+      ignoreDirs: ['ignored'],
+    });
+    expect(result).toEqual([targetFilePath]);
+  });
+
+  it('should respect the maxDirs limit and not find the file', async () => {
+    await createTestFile('content', 'a', 'b', 'c', 'target.txt');
+    const result = await bfsFileSearch(testRootDir, {
+      fileName: 'target.txt',
+      maxDirs: 3,
+    });
+    expect(result).toEqual([]);
+  });
+
+  it('should respect the maxDirs limit and find the file', async () => {
+    const targetFilePath = await createTestFile(
+      'content',
+      'a',
+      'b',
+      'c',
+      'target.txt',
+    );
+    const result = await bfsFileSearch(testRootDir, {
+      fileName: 'target.txt',
+      maxDirs: 4,
+    });
+    expect(result).toEqual([targetFilePath]);
+  });
+
+  describe('with FileDiscoveryService', () => {
+    let projectRoot: string;
+
+    beforeEach(async () => {
+      projectRoot = await createEmptyDir('project');
+    });
+
+    it('should ignore gitignored files', async () => {
+      await createEmptyDir('project', '.git');
+      await createTestFile('node_modules/', 'project', '.gitignore');
+      await createTestFile('content', 'project', 'node_modules', 'target.txt');
+      const targetFilePath = await createTestFile(
+        'content',
+        'project',
+        'not-ignored',
+        'target.txt',
+      );
+
+      const fileService = new FileDiscoveryService(projectRoot);
+      const result = await bfsFileSearch(projectRoot, {
+        fileName: 'target.txt',
+        fileService,
+        fileFilteringOptions: {
+          respectGitIgnore: true,
+          respectQwenIgnore: true,
+        },
+      });
+
+      expect(result).toEqual([targetFilePath]);
+    });
+
+    it('should ignore qwenignored files', async () => {
+      await createTestFile('node_modules/', 'project', '.qwenignore');
+      await createTestFile('content', 'project', 'node_modules', 'target.txt');
+      const targetFilePath = await createTestFile(
+        'content',
+        'project',
+        'not-ignored',
+        'target.txt',
+      );
+
+      const fileService = new FileDiscoveryService(projectRoot);
+      const result = await bfsFileSearch(projectRoot, {
+        fileName: 'target.txt',
+        fileService,
+        fileFilteringOptions: {
+          respectGitIgnore: false,
+          respectQwenIgnore: true,
+        },
+      });
+
+      expect(result).toEqual([targetFilePath]);
+    });
+
+    it('should not ignore files if respect flags are false', async () => {
+      await createEmptyDir('project', '.git');
+      await createTestFile('node_modules/', 'project', '.gitignore');
+      const target1 = await createTestFile(
+        'content',
+        'project',
+        'node_modules',
+        'target.txt',
+      );
+      const target2 = await createTestFile(
+        'content',
+        'project',
+        'not-ignored',
+        'target.txt',
+      );
+
+      const fileService = new FileDiscoveryService(projectRoot);
+      const result = await bfsFileSearch(projectRoot, {
+        fileName: 'target.txt',
+        fileService,
+        fileFilteringOptions: {
+          respectGitIgnore: false,
+          respectQwenIgnore: false,
+        },
+      });
+
+      expect(result.sort()).toEqual([target1, target2].sort());
+    });
+  });
+
+  it('should find all files in a complex directory structure', async () => {
+    // Create a complex directory structure to test correctness at scale
+    // without flaky performance checks.
+    const numDirs = 50;
+    const numFilesPerDir = 2;
+    const numTargetDirs = 10;
+
+    const dirCreationPromises: Array<Promise<unknown>> = [];
+    for (let i = 0; i < numDirs; i++) {
+      dirCreationPromises.push(createEmptyDir(`dir${i}`));
+      dirCreationPromises.push(createEmptyDir(`dir${i}`, 'subdir1'));
+      dirCreationPromises.push(createEmptyDir(`dir${i}`, 'subdir2'));
+      dirCreationPromises.push(createEmptyDir(`dir${i}`, 'subdir1', 'deep'));
+    }
+    await Promise.all(dirCreationPromises);
+
+    const fileCreationPromises: Array<Promise<string>> = [];
+    for (let i = 0; i < numTargetDirs; i++) {
+      // Add target files in some directories
+      fileCreationPromises.push(
+        createTestFile('content', `dir${i}`, 'QWEN.md'),
+      );
+      fileCreationPromises.push(
+        createTestFile('content', `dir${i}`, 'subdir1', 'QWEN.md'),
+      );
+    }
+    const expectedFiles = await Promise.all(fileCreationPromises);
+
+    const result = await bfsFileSearch(testRootDir, {
+      fileName: 'QWEN.md',
+      // Provide a generous maxDirs limit to ensure it doesn't prematurely stop
+      // in this large test case. Total dirs created is 200.
+      maxDirs: 250,
+    });
+
+    // Verify we found the exact files we created
+    expect(result.length).toBe(numTargetDirs * numFilesPerDir);
+    expect(result.sort()).toEqual(expectedFiles.sort());
+  });
+});
--- a/packages/core/src/utils/bfsFileSearch.ts
+++ b/packages/core/src/utils/bfsFileSearch.ts
@@ -0,0 +1,131 @@
+/**
+ * @license
+ * Copyright 2025 Google LLC
+ * SPDX-License-Identifier: Apache-2.0
+ */
+
+import * as fs from 'node:fs/promises';
+import * as path from 'node:path';
+import type { FileDiscoveryService } from '../services/fileDiscoveryService.js';
+import type { FileFilteringOptions } from '../config/constants.js';
+// Simple console logger for now.
+// TODO: Integrate with a more robust server-side logger.
+const logger = {
+  // eslint-disable-next-line @typescript-eslint/no-explicit-any
+  debug: (...args: any[]) => console.debug('[DEBUG] [BfsFileSearch]', ...args),
+};
+
+interface BfsFileSearchOptions {
+  fileName: string;
+  ignoreDirs?: string[];
+  maxDirs?: number;
+  debug?: boolean;
+  fileService?: FileDiscoveryService;
+  fileFilteringOptions?: FileFilteringOptions;
+}
+
+/**
+ * Performs a breadth-first search for a specific file within a directory structure.
+ *
+ * @param rootDir The directory to start the search from.
+ * @param options Configuration for the search.
+ * @returns A promise that resolves to an array of paths where the file was found.
+ */
+export async function bfsFileSearch(
+  rootDir: string,
+  options: BfsFileSearchOptions,
+): Promise<string[]> {
+  const {
+    fileName,
+    ignoreDirs = [],
+    maxDirs = Infinity,
+    debug = false,
+    fileService,
+  } = options;
+  const foundFiles: string[] = [];
+  const queue: string[] = [rootDir];
+  const visited = new Set<string>();
+  let scannedDirCount = 0;
+  let queueHead = 0; // Pointer-based queue head to avoid expensive splice operations
+
+  // Convert ignoreDirs array to Set for O(1) lookup performance
+  const ignoreDirsSet = new Set(ignoreDirs);
+
+  // Process directories in parallel batches for maximum performance
+  const PARALLEL_BATCH_SIZE = 15; // Parallel processing batch size for optimal performance
+
+  while (queueHead < queue.length && scannedDirCount < maxDirs) {
+    // Fill batch with unvisited directories up to the desired size
+    const batchSize = Math.min(PARALLEL_BATCH_SIZE, maxDirs - scannedDirCount);
+    const currentBatch = [];
+    while (currentBatch.length < batchSize && queueHead < queue.length) {
+      const currentDir = queue[queueHead];
+      queueHead++;
+      if (!visited.has(currentDir)) {
+        visited.add(currentDir);
+        currentBatch.push(currentDir);
+      }
+    }
+    scannedDirCount += currentBatch.length;
+
+    if (currentBatch.length === 0) continue;
+
+    if (debug) {
+      logger.debug(
+        `Scanning [${scannedDirCount}/${maxDirs}]: batch of ${currentBatch.length}`,
+      );
+    }
+
+    // Read directories in parallel instead of one by one
+    const readPromises = currentBatch.map(async (currentDir) => {
+      try {
+        const entries = await fs.readdir(currentDir, { withFileTypes: true });
+        return { currentDir, entries };
+      } catch (error) {
+        // Warn user that a directory could not be read, as this affects search results.
+        const message = (error as Error)?.message ?? 'Unknown error';
+        console.warn(
+          `[WARN] Skipping unreadable directory: ${currentDir} (${message})`,
+        );
+        if (debug) {
+          logger.debug(`Full error for ${currentDir}:`, error);
+        }
+        return { currentDir, entries: [] };
+      }
+    });
+
+    const results = await Promise.all(readPromises);
+
+    for (const { currentDir, entries } of results) {
+      for (const entry of entries) {
+        const fullPath = path.join(currentDir, entry.name);
+        const isDirectory = entry.isDirectory();
+        const isMatchingFile = entry.isFile() && entry.name === fileName;
+
+        if (!isDirectory && !isMatchingFile) {
+          continue;
+        }
+        if (isDirectory && ignoreDirsSet.has(entry.name)) {
+          continue;
+        }
+
+        if (
+          fileService?.shouldIgnoreFile(fullPath, {
+            respectGitIgnore: options.fileFilteringOptions?.respectGitIgnore,
+            respectQwenIgnore: options.fileFilteringOptions?.respectQwenIgnore,
+          })
+        ) {
+          continue;
+        }
+
+        if (isDirectory) {
+          queue.push(fullPath);
+        } else {
+          foundFiles.push(fullPath);
+        }
+      }
+    }
+  }
+
+  return foundFiles;
+}
--- a/packages/core/src/utils/fileUtils.test.ts
+++ b/packages/core/src/utils/fileUtils.test.ts
@@ -731,10 +731,6 @@ describe('fileUtils', () => {
      expect(
        (result.llmContent as { inlineData: { data: string } }).inlineData.data,
      ).toBe(fakePngData.toString('base64'));
-      expect(
-        (result.llmContent as { inlineData: { displayName?: string } })
-          .inlineData.displayName,
-      ).toBe('image.png');
      expect(result.returnDisplay).toContain('Read image file: image.png');
    });

@@ -747,20 +743,15 @@ describe('fileUtils', () => {
        mockConfig,
      );
      expect(
-        (result.llmContent as { fileData: unknown }).fileData,
+        (result.llmContent as { inlineData: unknown }).inlineData,
      ).toBeDefined();
      expect(
-        (result.llmContent as { fileData: { mimeType: string } }).fileData
+        (result.llmContent as { inlineData: { mimeType: string } }).inlineData
          .mimeType,
      ).toBe('application/pdf');
      expect(
-        (result.llmContent as { fileData: { fileUri: string } }).fileData
-          .fileUri,
+        (result.llmContent as { inlineData: { data: string } }).inlineData.data,
      ).toBe(fakePdfData.toString('base64'));
-      expect(
-        (result.llmContent as { fileData: { displayName?: string } }).fileData
-          .displayName,
-      ).toBe('document.pdf');
      expect(result.returnDisplay).toContain('Read pdf file: document.pdf');
    });

--- a/packages/core/src/utils/fileUtils.ts
+++ b/packages/core/src/utils/fileUtils.ts
@@ -351,7 +351,6 @@ export async function processSingleFileContent(
      .relative(rootDirectory, filePath)
      .replace(/\\/g, '/');

-    const displayName = path.basename(filePath);
    switch (fileType) {
      case 'binary': {
        return {
@@ -457,6 +456,7 @@ export async function processSingleFileContent(
        };
      }
      case 'image':
+      case 'pdf':
      case 'audio':
      case 'video': {
        const contentBuffer = await fs.promises.readFile(filePath);
@@ -466,21 +466,6 @@ export async function processSingleFileContent(
            inlineData: {
              data: base64Data,
              mimeType: mime.getType(filePath) || 'application/octet-stream',
-              displayName,
-            },
-          },
-          returnDisplay: `Read ${fileType} file: ${relativePathForDisplay}`,
-        };
-      }
-      case 'pdf': {
-        const contentBuffer = await fs.promises.readFile(filePath);
-        const base64Data = contentBuffer.toString('base64');
-        return {
-          llmContent: {
-            fileData: {
-              fileUri: base64Data,
-              mimeType: mime.getType(filePath) || 'application/octet-stream',
-              displayName,
            },
          },
          returnDisplay: `Read ${fileType} file: ${relativePathForDisplay}`,
--- a/packages/core/src/utils/memoryDiscovery.test.ts
+++ b/packages/core/src/utils/memoryDiscovery.test.ts
@@ -209,7 +209,7 @@ describe('loadServerHierarchicalMemory', () => {
    });
  });

-  it('should load context files from CWD with custom filename (not subdirectories)', async () => {
+  it('should load context files by downward traversal with custom filename', async () => {
    const customFilename = 'LOCAL_CONTEXT.md';
    setGeminiMdFilename(customFilename);

@@ -228,10 +228,9 @@ describe('loadServerHierarchicalMemory', () => {
      DEFAULT_FOLDER_TRUST,
    );

-    // Only upward traversal is performed, subdirectory files are not loaded
    expect(result).toEqual({
-      memoryContent: `--- Context from: ${customFilename} ---\nCWD custom memory\n--- End of Context from: ${customFilename} ---`,
-      fileCount: 1,
+      memoryContent: `--- Context from: ${customFilename} ---\nCWD custom memory\n--- End of Context from: ${customFilename} ---\n\n--- Context from: ${path.join('subdir', customFilename)} ---\nSubdir custom memory\n--- End of Context from: ${path.join('subdir', customFilename)} ---`,
+      fileCount: 2,
    });
  });

@@ -260,7 +259,7 @@ describe('loadServerHierarchicalMemory', () => {
    });
  });

-  it('should only load context files from CWD, not subdirectories', async () => {
+  it('should load ORIGINAL_GEMINI_MD_FILENAME files by downward traversal from CWD', async () => {
    await createTestFile(
      path.join(cwd, 'subdir', DEFAULT_CONTEXT_FILENAME),
      'Subdir memory',
@@ -279,14 +278,13 @@ describe('loadServerHierarchicalMemory', () => {
      DEFAULT_FOLDER_TRUST,
    );

-    // Subdirectory files are not loaded, only CWD and upward
    expect(result).toEqual({
-      memoryContent: `--- Context from: ${DEFAULT_CONTEXT_FILENAME} ---\nCWD memory\n--- End of Context from: ${DEFAULT_CONTEXT_FILENAME} ---`,
-      fileCount: 1,
+      memoryContent: `--- Context from: ${DEFAULT_CONTEXT_FILENAME} ---\nCWD memory\n--- End of Context from: ${DEFAULT_CONTEXT_FILENAME} ---\n\n--- Context from: ${path.join('subdir', DEFAULT_CONTEXT_FILENAME)} ---\nSubdir memory\n--- End of Context from: ${path.join('subdir', DEFAULT_CONTEXT_FILENAME)} ---`,
+      fileCount: 2,
    });
  });

-  it('should load and correctly order global and upward context files', async () => {
+  it('should load and correctly order global, upward, and downward ORIGINAL_GEMINI_MD_FILENAME files', async () => {
    const defaultContextFile = await createTestFile(
      path.join(homedir, QWEN_DIR, DEFAULT_CONTEXT_FILENAME),
      'default context content',
@@ -303,7 +301,7 @@ describe('loadServerHierarchicalMemory', () => {
      path.join(cwd, DEFAULT_CONTEXT_FILENAME),
      'CWD memory',
    );
-    await createTestFile(
+    const subDirGeminiFile = await createTestFile(
      path.join(cwd, 'sub', DEFAULT_CONTEXT_FILENAME),
      'Subdir memory',
    );
@@ -317,10 +315,92 @@ describe('loadServerHierarchicalMemory', () => {
      DEFAULT_FOLDER_TRUST,
    );

-    // Subdirectory files are not loaded, only global and upward from CWD
    expect(result).toEqual({
-      memoryContent: `--- Context from: ${path.relative(cwd, defaultContextFile)} ---\ndefault context content\n--- End of Context from: ${path.relative(cwd, defaultContextFile)} ---\n\n--- Context from: ${path.relative(cwd, rootGeminiFile)} ---\nProject parent memory\n--- End of Context from: ${path.relative(cwd, rootGeminiFile)} ---\n\n--- Context from: ${path.relative(cwd, projectRootGeminiFile)} ---\nProject root memory\n--- End of Context from: ${path.relative(cwd, projectRootGeminiFile)} ---\n\n--- Context from: ${path.relative(cwd, cwdGeminiFile)} ---\nCWD memory\n--- End of Context from: ${path.relative(cwd, cwdGeminiFile)} ---`,
-      fileCount: 4,
+      memoryContent: `--- Context from: ${path.relative(cwd, defaultContextFile)} ---\ndefault context content\n--- End of Context from: ${path.relative(cwd, defaultContextFile)} ---\n\n--- Context from: ${path.relative(cwd, rootGeminiFile)} ---\nProject parent memory\n--- End of Context from: ${path.relative(cwd, rootGeminiFile)} ---\n\n--- Context from: ${path.relative(cwd, projectRootGeminiFile)} ---\nProject root memory\n--- End of Context from: ${path.relative(cwd, projectRootGeminiFile)} ---\n\n--- Context from: ${path.relative(cwd, cwdGeminiFile)} ---\nCWD memory\n--- End of Context from: ${path.relative(cwd, cwdGeminiFile)} ---\n\n--- Context from: ${path.relative(cwd, subDirGeminiFile)} ---\nSubdir memory\n--- End of Context from: ${path.relative(cwd, subDirGeminiFile)} ---`,
+      fileCount: 5,
+    });
+  });
+
+  it('should ignore specified directories during downward scan', async () => {
+    await createEmptyDir(path.join(projectRoot, '.git'));
+    await createTestFile(path.join(projectRoot, '.gitignore'), 'node_modules');
+
+    await createTestFile(
+      path.join(cwd, 'node_modules', DEFAULT_CONTEXT_FILENAME),
+      'Ignored memory',
+    );
+    const regularSubDirGeminiFile = await createTestFile(
+      path.join(cwd, 'my_code', DEFAULT_CONTEXT_FILENAME),
+      'My code memory',
+    );
+
+    const result = await loadServerHierarchicalMemory(
+      cwd,
+      [],
+      false,
+      new FileDiscoveryService(projectRoot),
+      [],
+      DEFAULT_FOLDER_TRUST,
+      'tree',
+      {
+        respectGitIgnore: true,
+        respectQwenIgnore: true,
+      },
+      200, // maxDirs parameter
+    );
+
+    expect(result).toEqual({
+      memoryContent: `--- Context from: ${path.relative(cwd, regularSubDirGeminiFile)} ---\nMy code memory\n--- End of Context from: ${path.relative(cwd, regularSubDirGeminiFile)} ---`,
+      fileCount: 1,
+    });
+  });
+
+  it('should respect the maxDirs parameter during downward scan', async () => {
+    const consoleDebugSpy = vi
+      .spyOn(console, 'debug')
+      .mockImplementation(() => {});
+
+    // Create directories in parallel for better performance
+    const dirPromises = Array.from({ length: 2 }, (_, i) =>
+      createEmptyDir(path.join(cwd, `deep_dir_${i}`)),
+    );
+    await Promise.all(dirPromises);
+
+    // Pass the custom limit directly to the function
+    await loadServerHierarchicalMemory(
+      cwd,
+      [],
+      true,
+      new FileDiscoveryService(projectRoot),
+      [],
+      DEFAULT_FOLDER_TRUST,
+      'tree', // importFormat
+      {
+        respectGitIgnore: true,
+        respectQwenIgnore: true,
+      },
+      1, // maxDirs
+    );
+
+    expect(consoleDebugSpy).toHaveBeenCalledWith(
+      expect.stringContaining('[DEBUG] [BfsFileSearch]'),
+      expect.stringContaining('Scanning [1/1]:'),
+    );
+
+    vi.mocked(console.debug).mockRestore();
+
+    const result = await loadServerHierarchicalMemory(
+      cwd,
+      [],
+      false,
+      new FileDiscoveryService(projectRoot),
+      [],
+      DEFAULT_FOLDER_TRUST,
+    );
+
+    expect(result).toEqual({
+      memoryContent: '',
+      fileCount: 0,
    });
  });

--- a/packages/core/src/utils/memoryDiscovery.ts
+++ b/packages/core/src/utils/memoryDiscovery.ts
@@ -8,9 +8,12 @@ import * as fs from 'node:fs/promises';
 import * as fsSync from 'node:fs';
 import * as path from 'node:path';
 import { homedir } from 'node:os';
+import { bfsFileSearch } from './bfsFileSearch.js';
 import { getAllGeminiMdFilenames } from '../tools/memoryTool.js';
 import type { FileDiscoveryService } from '../services/fileDiscoveryService.js';
 import { processImports } from './memoryImportProcessor.js';
+import type { FileFilteringOptions } from '../config/constants.js';
+import { DEFAULT_MEMORY_FILE_FILTERING_OPTIONS } from '../config/constants.js';
 import { QWEN_DIR } from './paths.js';

 // Simple console logger, similar to the one previously in CLI's config.ts
@@ -83,6 +86,8 @@ async function getGeminiMdFilePathsInternal(
  fileService: FileDiscoveryService,
  extensionContextFilePaths: string[] = [],
  folderTrust: boolean,
+  fileFilteringOptions: FileFilteringOptions,
+  maxDirs: number,
 ): Promise<string[]> {
  const dirs = new Set<string>([
    ...includeDirectoriesToReadGemini,
@@ -104,6 +109,8 @@ async function getGeminiMdFilePathsInternal(
        fileService,
        extensionContextFilePaths,
        folderTrust,
+        fileFilteringOptions,
+        maxDirs,
      ),
    );

@@ -132,6 +139,8 @@ async function getGeminiMdFilePathsInternalForEachDir(
  fileService: FileDiscoveryService,
  extensionContextFilePaths: string[] = [],
  folderTrust: boolean,
+  fileFilteringOptions: FileFilteringOptions,
+  maxDirs: number,
 ): Promise<string[]> {
  const allPaths = new Set<string>();
  const geminiMdFilenames = getAllGeminiMdFilenames();
@@ -176,7 +185,7 @@ async function getGeminiMdFilePathsInternalForEachDir(
        // Not found, which is okay
      }
    } else if (dir && folderTrust) {
-      // FIX: Only perform the workspace search (upward scan from CWD to project root)
+      // FIX: Only perform the workspace search (upward and downward scans)
      // if a valid currentWorkingDirectory is provided and it's not the home directory.
      const resolvedCwd = path.resolve(dir);
      if (debugMode)
@@ -216,6 +225,23 @@ async function getGeminiMdFilePathsInternalForEachDir(
        currentDir = path.dirname(currentDir);
      }
      upwardPaths.forEach((p) => allPaths.add(p));
+
+      const mergedOptions: FileFilteringOptions = {
+        ...DEFAULT_MEMORY_FILE_FILTERING_OPTIONS,
+        ...fileFilteringOptions,
+      };
+
+      const downwardPaths = await bfsFileSearch(resolvedCwd, {
+        fileName: geminiMdFilename,
+        maxDirs,
+        debug: debugMode,
+        fileService,
+        fileFilteringOptions: mergedOptions,
+      });
+      downwardPaths.sort();
+      for (const dPath of downwardPaths) {
+        allPaths.add(dPath);
+      }
    }
  }

@@ -338,6 +364,8 @@ export async function loadServerHierarchicalMemory(
  extensionContextFilePaths: string[] = [],
  folderTrust: boolean,
  importFormat: 'flat' | 'tree' = 'tree',
+  fileFilteringOptions?: FileFilteringOptions,
+  maxDirs: number = 200,
 ): Promise<LoadServerHierarchicalMemoryResponse> {
  if (debugMode)
    logger.debug(
@@ -355,6 +383,8 @@ export async function loadServerHierarchicalMemory(
    fileService,
    extensionContextFilePaths,
    folderTrust,
+    fileFilteringOptions || DEFAULT_MEMORY_FILE_FILTERING_OPTIONS,
+    maxDirs,
  );
  if (filePaths.length === 0) {
    if (debugMode) logger.debug('No QWEN.md files found in hierarchy.');
@@ -370,14 +400,6 @@ export async function loadServerHierarchicalMemory(
    contentsWithPaths,
    currentWorkingDirectory,
  );
-
-  // Only count files that match configured memory filenames (e.g., QWEN.md),
-  // excluding system context files like output-language.md
-  const memoryFilenames = new Set(getAllGeminiMdFilenames());
-  const fileCount = contentsWithPaths.filter((item) =>
-    memoryFilenames.has(path.basename(item.filePath)),
-  ).length;
-
  if (debugMode)
    logger.debug(
      `Combined instructions length: ${combinedInstructions.length}`,
@@ -388,6 +410,6 @@ export async function loadServerHierarchicalMemory(
    );
  return {
    memoryContent: combinedInstructions,
-    fileCount, // Only count the context files
+    fileCount: contentsWithPaths.length,
  };
 }
--- a/packages/core/src/utils/pathReader.test.ts
+++ b/packages/core/src/utils/pathReader.test.ts
@@ -113,7 +113,6 @@ describe('readPathFromWorkspace', () => {
        inlineData: {
          mimeType: 'image/png',
          data: imageData.toString('base64'),
-          displayName: 'image.png',
        },
      },
    ]);
@@ -264,7 +263,6 @@ describe('readPathFromWorkspace', () => {
        inlineData: {
          mimeType: 'image/png',
          data: imageData.toString('base64'),
-          displayName: 'photo.png',
        },
      });
    });
--- a/packages/test-utils/package.json
+++ b/packages/test-utils/package.json
@@ -1,6 +1,6 @@
 {
  "name": "@qwen-code/qwen-code-test-utils",
-  "version": "0.7.1",
+  "version": "0.7.0",
  "private": true,
  "main": "src/index.ts",
  "license": "Apache-2.0",
--- a/packages/vscode-ide-companion/README.md
+++ b/packages/vscode-ide-companion/README.md
@@ -1,11 +1,6 @@
 # Qwen Code Companion

-[![Version](https://img.shields.io/visual-studio-marketplace/v/qwenlm.qwen-code-vscode-ide-companion)](https://marketplace.visualstudio.com/items?itemName=qwenlm.qwen-code-vscode-ide-companion)
-[![VS Code Installs](https://img.shields.io/visual-studio-marketplace/i/qwenlm.qwen-code-vscode-ide-companion)](https://marketplace.visualstudio.com/items?itemName=qwenlm.qwen-code-vscode-ide-companion)
-[![Open VSX Downloads](https://img.shields.io/open-vsx/dt/qwenlm/qwen-code-vscode-ide-companion)](https://open-vsx.org/extension/qwenlm/qwen-code-vscode-ide-companion)
-[![Rating](https://img.shields.io/visual-studio-marketplace/r/qwenlm.qwen-code-vscode-ide-companion)](https://marketplace.visualstudio.com/items?itemName=qwenlm.qwen-code-vscode-ide-companion)
-
-Seamlessly integrate [Qwen Code](https://github.com/QwenLM/qwen-code) into Visual Studio Code with native IDE features and an intuitive chat interface. This extension bundles everything you need — no additional installation required.
+Seamlessly integrate [Qwen Code](https://github.com/QwenLM/qwen-code) into Visual Studio Code with native IDE features and an intuitive interface. This extension bundles everything you need to get started immediately.

 ## Demo

@@ -16,7 +11,7 @@ Seamlessly integrate [Qwen Code](https://github.com/QwenLM/qwen-code) into Visua

 ## Features

- **Native IDE experience**: Dedicated Qwen Code Chat panel accessed via the Qwen icon in the editor title bar
+- **Native IDE experience**: Dedicated Qwen Code sidebar panel accessed via the Qwen icon
 - **Native diffing**: Review, edit, and accept changes in VS Code's diff view
 - **Auto-accept edits mode**: Automatically apply Qwen's changes as they're made
 - **File management**: @-mention files or attach files and images using the system file picker
@@ -25,46 +20,73 @@ Seamlessly integrate [Qwen Code](https://github.com/QwenLM/qwen-code) into Visua

 ## Requirements

- Visual Studio Code 1.85.0 or newer (also works with Cursor, Windsurf, and other VS Code-based editors)
+- Visual Studio Code 1.85.0 or newer

-## Quick Start
+## Installation

-1. **Install** from the [VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=qwenlm.qwen-code-vscode-ide-companion) or [Open VSX Registry](https://open-vsx.org/extension/qwenlm/qwen-code-vscode-ide-companion)
+1. Install from the VS Code Marketplace: https://marketplace.visualstudio.com/items?itemName=qwenlm.qwen-code-vscode-ide-companion

-2. **Open the Chat panel** using one of these methods:
-   - Click the **Qwen icon** in the top-right corner of the editor
-   - Run `Qwen Code: Open` from the Command Palette (`Cmd+Shift+P` / `Ctrl+Shift+P`)
+2. Two ways to use
+   - Chat panel: Click the Qwen icon in the Activity Bar, or run `Qwen Code: Open` from the Command Palette (`Cmd+Shift+P` / `Ctrl+Shift+P`).
+   - Terminal session (classic): Run `Qwen Code: Run` to launch a session in the integrated terminal (bundled CLI).

-3. **Start chatting** — Ask Qwen to help with coding tasks, explain code, fix bugs, or write new features
+## Development and Debugging

-## Commands
+To debug and develop this extension locally:

-| Command                          | Description                                            |
-| -------------------------------- | ------------------------------------------------------ |
-| `Qwen Code: Open`                | Open the Qwen Code Chat panel                          |
-| `Qwen Code: Run`                 | Launch a classic terminal session with the bundled CLI |
-| `Qwen Code: Accept Current Diff` | Accept the currently displayed diff                    |
-| `Qwen Code: Close Diff Editor`   | Close/reject the current diff                          |
+1. **Clone the repository**

-## Feedback & Issues
+   ```bash
+   git clone https://github.com/QwenLM/qwen-code.git
+   cd qwen-code
+   ```

- 🐛 [Report bugs](https://github.com/QwenLM/qwen-code/issues/new?template=bug_report.yml&labels=bug,vscode-ide-companion)
- 💡 [Request features](https://github.com/QwenLM/qwen-code/issues/new?template=feature_request.yml&labels=enhancement,vscode-ide-companion)
- 📖 [Documentation](https://qwenlm.github.io/qwen-code-docs/)
- 📋 [Changelog](https://github.com/QwenLM/qwen-code/releases)
+2. **Install dependencies**

-## Contributing
+   ```bash
+   npm install
+   # or if using pnpm
+   pnpm install
+   ```

-We welcome contributions! See our [Contributing Guide](https://github.com/QwenLM/qwen-code/blob/main/CONTRIBUTING.md) for details on:
+3. **Start debugging**

- Setting up the development environment
- Building and debugging the extension locally
- Submitting pull requests
+   ```bash
+   code .  # Open the project root in VS Code
+   ```
+   - Open the `packages/vscode-ide-companion/src/extension.ts` file
+   - Open Debug panel (`Ctrl+Shift+D` or `Cmd+Shift+D`)
+   - Select **"Launch Companion VS Code Extension"** from the debug dropdown
+   - Press `F5` to launch Extension Development Host
+
+4. **Make changes and reload**
+   - Edit the source code in the original VS Code window
+   - To see your changes, reload the Extension Development Host window by:
+     - Pressing `Ctrl+R` (Windows/Linux) or `Cmd+R` (macOS)
+     - Or clicking the "Reload" button in the debug toolbar
+
+5. **View logs and debug output**
+   - Open the Debug Console in the original VS Code window to see extension logs
+   - In the Extension Development Host window, open Developer Tools with `Help > Toggle Developer Tools` to see webview logs
+
+## Build for Production
+
+To build the extension for distribution:
+
+```bash
+npm run compile
+# or
+pnpm run compile
+```
+
+To package the extension as a VSIX file:
+
+```bash
+npx vsce package
+# or
+pnpm vsce package
+```

 ## Terms of Service and Privacy Notice

 By installing this extension, you agree to the [Terms of Service](https://github.com/QwenLM/qwen-code/blob/main/docs/tos-privacy.md).
-
-## License
-
-[Apache-2.0](https://github.com/QwenLM/qwen-code/blob/main/LICENSE)
--- a/packages/vscode-ide-companion/package.json
+++ b/packages/vscode-ide-companion/package.json
@@ -2,7 +2,7 @@
  "name": "qwen-code-vscode-ide-companion",
  "displayName": "Qwen Code Companion",
  "description": "Enable Qwen Code with direct access to your VS Code workspace.",
-  "version": "0.7.1",
+  "version": "0.7.0",
  "publisher": "qwenlm",
  "icon": "assets/icon.png",
  "repository": {
--- a/packages/vscode-ide-companion/src/constants/acpSchema.ts
+++ b/packages/vscode-ide-companion/src/constants/acpSchema.ts
@@ -23,23 +23,3 @@ export const CLIENT_METHODS = {
  session_request_permission: 'session/request_permission',
  session_update: 'session/update',
 } as const;
-
-export const ACP_ERROR_CODES = {
-  // Parse error: invalid JSON received by server.
-  PARSE_ERROR: -32700,
-  // Invalid request: JSON is not a valid Request object.
-  INVALID_REQUEST: -32600,
-  // Method not found: method does not exist or is unavailable.
-  METHOD_NOT_FOUND: -32601,
-  // Invalid params: invalid method parameter(s).
-  INVALID_PARAMS: -32602,
-  // Internal error: implementation-defined server error.
-  INTERNAL_ERROR: -32603,
-  // Authentication required: must authenticate before operation.
-  AUTH_REQUIRED: -32000,
-  // Resource not found: e.g. missing file.
-  RESOURCE_NOT_FOUND: -32002,
-} as const;
-
-export type AcpErrorCode =
-  (typeof ACP_ERROR_CODES)[keyof typeof ACP_ERROR_CODES];
--- a/packages/vscode-ide-companion/src/extension.ts
+++ b/packages/vscode-ide-companion/src/extension.ts
@@ -314,32 +314,34 @@ export async function activate(context: vscode.ExtensionContext) {
            'cli.js',
          ).fsPath;
          const execPath = process.execPath;
+          const lowerExecPath = execPath.toLowerCase();
+          const needsElectronRunAsNode =
+            lowerExecPath.includes('code') ||
+            lowerExecPath.includes('electron');

+          let qwenCmd: string;
          const terminalOptions: vscode.TerminalOptions = {
            name: `Qwen Code (${selectedFolder.name})`,
            cwd: selectedFolder.uri.fsPath,
            location,
          };

-          let qwenCmd: string;
-
          if (isWindows) {
-            // On Windows, try multiple strategies to find a Node.js runtime:
-            // 1. Check if VSCode ships a standalone node.exe alongside Code.exe
-            // 2. Check VSCode's internal Node.js in resources directory
-            // 3. Fall back to using Code.exe with ELECTRON_RUN_AS_NODE=1
+            // Use system Node via cmd.exe; avoid PowerShell parsing issues
            const quoteCmd = (s: string) => `"${s.replace(/"/g, '""')}"`;
            const cliQuoted = quoteCmd(cliEntry);
            // TODO: @yiliang114, temporarily run through node, and later hope to decouple from the local node
            qwenCmd = `node ${cliQuoted}`;
            terminalOptions.shellPath = process.env.ComSpec;
          } else {
-            // macOS/Linux: All VSCode-like IDEs (VSCode, Cursor, Windsurf, etc.)
-            // are Electron-based, so we always need ELECTRON_RUN_AS_NODE=1
-            // to run Node.js scripts using the IDE's bundled runtime.
            const quotePosix = (s: string) => `"${s.replace(/"/g, '\\"')}"`;
            const baseCmd = `${quotePosix(execPath)} ${quotePosix(cliEntry)}`;
-            qwenCmd = `ELECTRON_RUN_AS_NODE=1 ${baseCmd}`;
+            if (needsElectronRunAsNode) {
+              // macOS Electron helper needs ELECTRON_RUN_AS_NODE=1;
+              qwenCmd = `ELECTRON_RUN_AS_NODE=1 ${baseCmd}`;
+            } else {
+              qwenCmd = baseCmd;
+            }
          }

          const terminal = vscode.window.createTerminal(terminalOptions);
--- a/packages/vscode-ide-companion/src/ide-server.ts
+++ b/packages/vscode-ide-companion/src/ide-server.ts
@@ -28,7 +28,6 @@ import * as os from 'node:os';
 import type { z } from 'zod';
 import type { DiffManager } from './diff-manager.js';
 import { OpenFilesManager } from './open-files-manager.js';
-import { ACP_ERROR_CODES } from './constants/acpSchema.js';

 class CORSError extends Error {
  constructor(message: string) {
@@ -265,7 +264,7 @@ export class IDEServer {
          res.status(400).json({
            jsonrpc: '2.0',
            error: {
-              code: ACP_ERROR_CODES.AUTH_REQUIRED,
+              code: -32000,
              message:
                'Bad Request: No valid session ID provided for non-initialize request.',
            },
@@ -284,7 +283,7 @@ export class IDEServer {
            res.status(500).json({
              jsonrpc: '2.0' as const,
              error: {
-                code: ACP_ERROR_CODES.INTERNAL_ERROR,
+                code: -32603,
                message: 'Internal server error',
              },
              id: null,
--- a/packages/vscode-ide-companion/src/services/acpConnection.ts
+++ b/packages/vscode-ide-companion/src/services/acpConnection.ts
@@ -5,7 +5,6 @@
 */

 import { JSONRPC_VERSION } from '../types/acpTypes.js';
-import { ACP_ERROR_CODES } from '../constants/acpSchema.js';
 import type {
  AcpMessage,
  AcpPermissionRequest,
@@ -233,34 +232,12 @@ export class AcpConnection {
        })
        .catch((error) => {
          if ('id' in message && typeof message.id === 'number') {
-            const errorMessage =
-              error instanceof Error
-                ? error.message
-                : typeof error === 'object' &&
-                    error !== null &&
-                    'message' in error &&
-                    typeof (error as { message: unknown }).message === 'string'
-                  ? (error as { message: string }).message
-                  : String(error);
-
-            let errorCode: number = ACP_ERROR_CODES.INTERNAL_ERROR;
-            const errorCodeValue =
-              typeof error === 'object' && error !== null && 'code' in error
-                ? (error as { code?: unknown }).code
-                : undefined;
-
-            if (typeof errorCodeValue === 'number') {
-              errorCode = errorCodeValue;
-            } else if (errorCodeValue === 'ENOENT') {
-              errorCode = ACP_ERROR_CODES.RESOURCE_NOT_FOUND;
-            }
-
            this.messageHandler.sendResponseMessage(this.child, {
              jsonrpc: JSONRPC_VERSION,
              id: message.id,
              error: {
-                code: errorCode,
-                message: errorMessage,
+                code: -32603,
+                message: error instanceof Error ? error.message : String(error),
              },
            });
          }
--- a/packages/vscode-ide-companion/src/services/acpFileHandler.ts
+++ b/packages/vscode-ide-companion/src/services/acpFileHandler.ts
@@ -66,11 +66,6 @@ export class AcpFileHandler {
      const errorMsg = error instanceof Error ? error.message : String(error);
      console.error(`[ACP] Failed to read file ${params.path}:`, errorMsg);

-      const nodeError = error as NodeJS.ErrnoException;
-      if (nodeError?.code === 'ENOENT') {
-        throw error;
-      }
-
      throw new Error(`Failed to read file '${params.path}': ${errorMsg}`);
    }
  }
--- a/packages/vscode-ide-companion/src/utils/authErrors.ts
+++ b/packages/vscode-ide-companion/src/utils/authErrors.ts
@@ -4,11 +4,9 @@
 * SPDX-License-Identifier: Apache-2.0
 */

-import { ACP_ERROR_CODES } from '../constants/acpSchema.js';
-
 const AUTH_ERROR_PATTERNS = [
  'Authentication required', // Standard authentication request message
-  `(code: ${ACP_ERROR_CODES.AUTH_REQUIRED})`, // RPC error code indicates auth failure
+  '(code: -32000)', // RPC error code -32000 indicates authentication failure
  'Unauthorized', // HTTP unauthorized error
  'Invalid token', // Invalid token
  'Session expired', // Session expired
--- a/packages/vscode-ide-companion/src/webview/handlers/SessionMessageHandler.ts
+++ b/packages/vscode-ide-companion/src/webview/handlers/SessionMessageHandler.ts
@@ -8,9 +8,6 @@ import * as vscode from 'vscode';
 import { BaseMessageHandler } from './BaseMessageHandler.js';
 import type { ChatMessage } from '../../services/qwenAgentManager.js';
 import type { ApprovalModeValue } from '../../types/approvalModeValueTypes.js';
-import { ACP_ERROR_CODES } from '../../constants/acpSchema.js';
-
-const AUTH_REQUIRED_CODE_PATTERN = `(code: ${ACP_ERROR_CODES.AUTH_REQUIRED})`;

 /**
 * Session message handler
@@ -358,7 +355,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
          createErr instanceof Error ? createErr.message : String(createErr);
        if (
          errorMsg.includes('Authentication required') ||
-          errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN)
+          errorMsg.includes('(code: -32000)')
        ) {
          await this.promptLogin(
            'Your login session has expired or is invalid. Please login again to continue using Qwen Code.',
@@ -424,7 +421,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
        errorMsg.includes('Session not found') ||
        errorMsg.includes('No active ACP session') ||
        errorMsg.includes('Authentication required') ||
-        errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+        errorMsg.includes('(code: -32000)') ||
        errorMsg.includes('Unauthorized') ||
        errorMsg.includes('Invalid token')
      ) {
@@ -515,7 +512,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
      // Check for authentication/session expiration errors
      if (
        errorMsg.includes('Authentication required') ||
-        errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+        errorMsg.includes('(code: -32000)') ||
        errorMsg.includes('Unauthorized') ||
        errorMsg.includes('Invalid token') ||
        errorMsg.includes('No active ACP session')
@@ -625,7 +622,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
        // Check for authentication/session expiration errors
        if (
          errorMsg.includes('Authentication required') ||
-          errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+          errorMsg.includes('(code: -32000)') ||
          errorMsg.includes('Unauthorized') ||
          errorMsg.includes('Invalid token') ||
          errorMsg.includes('No active ACP session')
@@ -685,7 +682,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
            // Check for authentication/session expiration errors in session creation
            if (
              createErrorMsg.includes('Authentication required') ||
-              createErrorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+              createErrorMsg.includes('(code: -32000)') ||
              createErrorMsg.includes('Unauthorized') ||
              createErrorMsg.includes('Invalid token') ||
              createErrorMsg.includes('No active ACP session')
@@ -725,7 +722,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
      // Check for authentication/session expiration errors
      if (
        errorMsg.includes('Authentication required') ||
-        errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+        errorMsg.includes('(code: -32000)') ||
        errorMsg.includes('Unauthorized') ||
        errorMsg.includes('Invalid token') ||
        errorMsg.includes('No active ACP session')
@@ -780,7 +777,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
      // Check for authentication/session expiration errors
      if (
        errorMsg.includes('Authentication required') ||
-        errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+        errorMsg.includes('(code: -32000)') ||
        errorMsg.includes('Unauthorized') ||
        errorMsg.includes('Invalid token') ||
        errorMsg.includes('No active ACP session')
@@ -830,7 +827,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
        // Check for authentication/session expiration errors
        if (
          errorMsg.includes('Authentication required') ||
-          errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+          errorMsg.includes('(code: -32000)') ||
          errorMsg.includes('Unauthorized') ||
          errorMsg.includes('Invalid token') ||
          errorMsg.includes('No active ACP session')
@@ -858,7 +855,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
      // Check for authentication/session expiration errors
      if (
        errorMsg.includes('Authentication required') ||
-        errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+        errorMsg.includes('(code: -32000)') ||
        errorMsg.includes('Unauthorized') ||
        errorMsg.includes('Invalid token') ||
        errorMsg.includes('No active ACP session')
@@ -964,7 +961,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
        // Check for authentication/session expiration errors
        if (
          errorMsg.includes('Authentication required') ||
-          errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+          errorMsg.includes('(code: -32000)') ||
          errorMsg.includes('Unauthorized') ||
          errorMsg.includes('Invalid token') ||
          errorMsg.includes('No active ACP session')
@@ -992,7 +989,7 @@ export class SessionMessageHandler extends BaseMessageHandler {
      // Check for authentication/session expiration errors
      if (
        errorMsg.includes('Authentication required') ||
-        errorMsg.includes(AUTH_REQUIRED_CODE_PATTERN) ||
+        errorMsg.includes('(code: -32000)') ||
        errorMsg.includes('Unauthorized') ||
        errorMsg.includes('Invalid token') ||
        errorMsg.includes('No active ACP session')