feihong-code 7.2.0 → 7.5.0

package/docs/UPDATE_NOTES_en.md

@@ -1,159 +1,159 @@
-# Feihong Code v7.0.0 Release Notes
-
-**Release Date**: August 24, 2026
-**Version**: 7.0.0
-**Product Name**: Feihong Code (feihong-code)
-**R&D Team**: Feihongzhi Technology Enterprise Management Co., Ltd. · Feiyang Qiyuan R&D Center
-
----
-
-## Version Overview
-
-Feihong Code v7.0.0 is a major version upgrade, jumping directly from 0.6.1 to 7.0.0, marking the product's mature transformation from a terminal tool to a full-platform AI coding assistant. This release focuses on optimizing user experience, adding desktop support, improving multilingual internationalization, and significantly enhancing the naturalness of conversational interaction.
-
----
-
-## Key Highlights
-
-### 1. Complete Conversation Flow Refactoring (Doubao-like Experience)
-
-- **Pure Text Output**: Completely removed Markdown formatting, tool call status, technical summaries, and other distracting information
-- **Real-time Thinking Process**: Model's thinking process displayed in real-time, letting users see what the AI is thinking
-- **Conversational Tone**: Natural like human conversation, no longer technical report style
-- **Hidden Operation Steps**: No longer shows "Executing operation: read_file, run_shell" and other technical details
-- **Separation of Thinking and Reply**: Thinking process displayed in real-time, final reply concise and clear, no repetition
-
-### 2. Electron Desktop Version Officially Released
-
-- **Independent Desktop Window**: No longer dependent on browser, double-click to use
-- **System Tray Support**: Minimize to background, invoke anytime
-- **Single Instance Prevention**: Single instance operation, avoiding resource waste
-- **Application Menu**: Complete edit menu (copy/paste/cut/select all)
-- **Permission Management**: Microphone, camera, screen capture permissions automatically handled
-- **One-click Startup Script**: `一键启动桌面版.bat`, auto build + launch
-
-### 3. Interface and Interaction Optimization
-
-- **Interface Width Optimization**: Main area max width 1400px, no longer fills the entire screen
-- **User Menu Refactored**: Changed from bottom pull-up to right sidebar (280px)
-- **Startup Logic Fixed**: No longer opens two web pages, one launch directly enters
-- **Text Selection Copy**: Support selecting partial text for copying, no longer only whole message copy
-- **Message Action Buttons**: Hover to show copy, create document buttons
-- **Conversation History Expansion**: conversation limit increased from 80 to 300, steps limit from 200 to 500
-
-### 4. New and Improved Features
-
-- **Screenshot Function**: Support screen capture + crop selection, screenshot directly to input box
-- **Computer Control**: Backend PowerShell API, support screenshot, mouse move/click, keyboard input
-- **Voice Input**: Web Speech API, recommended to use Edge browser
-- **Voice/Video Call**: Real-time microphone recognition + camera view + video snapshot
-- **Workspace Picker**: Drives and folders paginated, support create/rename folders
-- **Conversation Task Optimization**: Click conversation directly creates new task, no need to manually click new
-- **Automation Quick Commands**: Built-in common quick commands
-- **Template Library Plugin System**: Support custom sources, nodes connect to external plugins
-- **Memory System**: Long-term memory customization, add record simplified to plain text input
-
-### 5. Enhanced Coding and Bug Fixing Capabilities
-
-- **System Prompt Upgrade**: Added 8 coding capability requirements, 7 bug fixing capability requirements
-- **Error Classification Optimization**: build-error and other errors provide 5-step specific troubleshooting process
-- **Continuous Error Handling**: Traverse all continuous errors for classification recording, no longer only record the last one
-- **Self-check Maintenance Optimization**: Only the first task per day can do environment confirmation, subsequent tasks execute directly, only fix when errors occur
-- **Build Error Self-healing**: Structured report on failure (error type + last error + fix suggestion + next step)
-
-### 6. Chinese-English Internationalization
-
-- **Backend i18n**: Complete Chinese-English entries
-- **Frontend Internationalization**: Computer control, video call, message operations, etc. added Chinese-English translations
-- **Continuously Improving**: Some hardcoded text gradually migrated to i18n
-
-### 7. Local Speech Recognition Service (Optional)
-
-- **faster-whisper Local Deployment**: Completely free, offline, no call limits
-- **HTTP API Interface**: Any project can call
-- **Multilingual Support**: Chinese optimized, high accuracy
-- **Independent Service**: Not coupled with main program, start on demand
-
----
-
-## Bug Fixes
-
-- Fixed Electron desktop version auto-exit issue
-- Fixed issue of opening two web pages after startup
-- Fixed conversation history too long causing earlier content to be truncated
-- Fixed thinking process not displaying in real-time during task execution
-- Fixed desktop version copy-paste shortcut failure
-- Fixed desktop version screenshot permission repeated request
-- Fixed 13389.js s.nodeName.toLowerCase error (non-project code, confirmed not affecting functionality)
-- Fixed inaccurate error information after 3 consecutive build errors
-
----
-
-## Upgrade Guide
-
-### Upgrade from 0.6.1
-
-```bash
-# Pull latest code
-git pull origin master
-
-# Install dependencies (if any new)
-npm install
-
-# Rebuild
-npm run build
-
-# Start service
-fhcode serve
-# or
-node dist/cli/index.js serve
-```
-
-### Desktop Version Usage
-
-```bash
-# Method 1: One-click startup
-Double-click 一键启动桌面版.bat
-
-# Method 2: Manual startup
-npm run build
-npm run electron
-```
-
-### Local Speech Recognition Service (Optional)
-
-```bash
-cd funasr-server
-Double-click 启动语音识别服务.bat
-# Service starts at http://localhost:8082
-```
-
----
-
-## System Requirements
-
-- **Node.js**: 18.0 or higher
-- **Operating System**: Windows 10/11 (recommended), macOS, Linux
-- **Memory**: At least 4GB (8GB recommended)
-- **Desktop Version**: Windows 10/11, Electron v41+
-
----
-
-## Known Limitations
-
-- Desktop version voice input depends on Google services, domestically recommended to use web version (Edge browser) or local speech recognition service
-- Computer control feature currently manual operation interface, AI automatic operation under development
-- Chinese-English internationalization not yet fully covering all hardcoded text
-- Local speech recognition service is non-real-time (recognizes after recording completes)
-
----
-
-## Feedback and Support
-
-- GitHub Issues: https://github.com/wch887292/feihong-code/issues
-- npm: https://www.npmjs.com/package/feihong-code
-
----
-
-*Feihongzhi Technology Enterprise Management Co., Ltd. · Feiyang Qiyuan R&D Center*
-*Person in Charge: Wu Cihong*
+# Feihong Code v7.0.0 Release Notes
+
+**Release Date**: August 24, 2026
+**Version**: 7.0.0
+**Product Name**: Feihong Code (feihong-code)
+**R&D Team**: Feihongzhi Technology Enterprise Management Co., Ltd. · Feiyang Qiyuan R&D Center
+
+---
+
+## Version Overview
+
+Feihong Code v7.0.0 is a major version upgrade, jumping directly from 0.6.1 to 7.0.0, marking the product's mature transformation from a terminal tool to a full-platform AI coding assistant. This release focuses on optimizing user experience, adding desktop support, improving multilingual internationalization, and significantly enhancing the naturalness of conversational interaction.
+
+---
+
+## Key Highlights
+
+### 1. Complete Conversation Flow Refactoring (Doubao-like Experience)
+
+- **Pure Text Output**: Completely removed Markdown formatting, tool call status, technical summaries, and other distracting information
+- **Real-time Thinking Process**: Model's thinking process displayed in real-time, letting users see what the AI is thinking
+- **Conversational Tone**: Natural like human conversation, no longer technical report style
+- **Hidden Operation Steps**: No longer shows "Executing operation: read_file, run_shell" and other technical details
+- **Separation of Thinking and Reply**: Thinking process displayed in real-time, final reply concise and clear, no repetition
+
+### 2. Electron Desktop Version Officially Released
+
+- **Independent Desktop Window**: No longer dependent on browser, double-click to use
+- **System Tray Support**: Minimize to background, invoke anytime
+- **Single Instance Prevention**: Single instance operation, avoiding resource waste
+- **Application Menu**: Complete edit menu (copy/paste/cut/select all)
+- **Permission Management**: Microphone, camera, screen capture permissions automatically handled
+- **One-click Startup Script**: `一键启动桌面版.bat`, auto build + launch
+
+### 3. Interface and Interaction Optimization
+
+- **Interface Width Optimization**: Main area max width 1400px, no longer fills the entire screen
+- **User Menu Refactored**: Changed from bottom pull-up to right sidebar (280px)
+- **Startup Logic Fixed**: No longer opens two web pages, one launch directly enters
+- **Text Selection Copy**: Support selecting partial text for copying, no longer only whole message copy
+- **Message Action Buttons**: Hover to show copy, create document buttons
+- **Conversation History Expansion**: conversation limit increased from 80 to 300, steps limit from 200 to 500
+
+### 4. New and Improved Features
+
+- **Screenshot Function**: Support screen capture + crop selection, screenshot directly to input box
+- **Computer Control**: Backend PowerShell API, support screenshot, mouse move/click, keyboard input
+- **Voice Input**: Web Speech API, recommended to use Edge browser
+- **Voice/Video Call**: Real-time microphone recognition + camera view + video snapshot
+- **Workspace Picker**: Drives and folders paginated, support create/rename folders
+- **Conversation Task Optimization**: Click conversation directly creates new task, no need to manually click new
+- **Automation Quick Commands**: Built-in common quick commands
+- **Template Library Plugin System**: Support custom sources, nodes connect to external plugins
+- **Memory System**: Long-term memory customization, add record simplified to plain text input
+
+### 5. Enhanced Coding and Bug Fixing Capabilities
+
+- **System Prompt Upgrade**: Added 8 coding capability requirements, 7 bug fixing capability requirements
+- **Error Classification Optimization**: build-error and other errors provide 5-step specific troubleshooting process
+- **Continuous Error Handling**: Traverse all continuous errors for classification recording, no longer only record the last one
+- **Self-check Maintenance Optimization**: Only the first task per day can do environment confirmation, subsequent tasks execute directly, only fix when errors occur
+- **Build Error Self-healing**: Structured report on failure (error type + last error + fix suggestion + next step)
+
+### 6. Chinese-English Internationalization
+
+- **Backend i18n**: Complete Chinese-English entries
+- **Frontend Internationalization**: Computer control, video call, message operations, etc. added Chinese-English translations
+- **Continuously Improving**: Some hardcoded text gradually migrated to i18n
+
+### 7. Local Speech Recognition Service (Optional)
+
+- **faster-whisper Local Deployment**: Completely free, offline, no call limits
+- **HTTP API Interface**: Any project can call
+- **Multilingual Support**: Chinese optimized, high accuracy
+- **Independent Service**: Not coupled with main program, start on demand
+
+---
+
+## Bug Fixes
+
+- Fixed Electron desktop version auto-exit issue
+- Fixed issue of opening two web pages after startup
+- Fixed conversation history too long causing earlier content to be truncated
+- Fixed thinking process not displaying in real-time during task execution
+- Fixed desktop version copy-paste shortcut failure
+- Fixed desktop version screenshot permission repeated request
+- Fixed 13389.js s.nodeName.toLowerCase error (non-project code, confirmed not affecting functionality)
+- Fixed inaccurate error information after 3 consecutive build errors
+
+---
+
+## Upgrade Guide
+
+### Upgrade from 0.6.1
+
+```bash
+# Pull latest code
+git pull origin master
+
+# Install dependencies (if any new)
+npm install
+
+# Rebuild
+npm run build
+
+# Start service
+fhcode serve
+# or
+node dist/cli/index.js serve
+```
+
+### Desktop Version Usage
+
+```bash
+# Method 1: One-click startup
+Double-click 一键启动桌面版.bat
+
+# Method 2: Manual startup
+npm run build
+npm run electron
+```
+
+### Local Speech Recognition Service (Optional)
+
+```bash
+cd funasr-server
+Double-click 启动语音识别服务.bat
+# Service starts at http://localhost:8082
+```
+
+---
+
+## System Requirements
+
+- **Node.js**: 18.0 or higher
+- **Operating System**: Windows 10/11 (recommended), macOS, Linux
+- **Memory**: At least 4GB (8GB recommended)
+- **Desktop Version**: Windows 10/11, Electron v41+
+
+---
+
+## Known Limitations
+
+- Desktop version voice input depends on Google services, domestically recommended to use web version (Edge browser) or local speech recognition service
+- Computer control feature currently manual operation interface, AI automatic operation under development
+- Chinese-English internationalization not yet fully covering all hardcoded text
+- Local speech recognition service is non-real-time (recognizes after recording completes)
+
+---
+
+## Feedback and Support
+
+- GitHub Issues: https://github.com/wch887292/feihong-code/issues
+- npm: https://www.npmjs.com/package/feihong-code
+
+---
+
+*Feihongzhi Technology Enterprise Management Co., Ltd. · Feiyang Qiyuan R&D Center*
+*Person in Charge: Wu Cihong*

package/docs/UPDATE_NOTES_zh.md

@@ -1,159 +1,159 @@
-# 飞虹 Code v7.0.0 更新说明书
-
-**发布日期**：2026年8月24日
-**版本号**：7.0.0
-**产品名称**：飞虹 Code（feihong-code）
-**研发团队**：晋江市飞虹智科技企业管理有限公司 · 飞扬企源研发中心
-
----
-
-## 版本概述
-
-飞虹 Code v7.0.0 是一个重大版本升级，从 0.6.1 直接跨越到 7.0.0，标志着产品从终端工具向全平台 AI 编程助手的成熟蜕变。本版本重点优化了用户体验、新增了桌面端支持、完善了多语言国际化，并大幅提升了对话交互的自然度。
-
----
-
-## 核心亮点
-
-### 1. 对话流全面重构（对标豆包体验）
-
-- **纯文本输出**：彻底去除 Markdown 格式、工具调用状态、技术摘要等干扰信息
-- **实时思考过程**：模型思考过程实时展示，让用户看到 AI 在想什么
-- **口语化表达**：像普通人交流一样自然，不再是技术报告风格
-- **操作步骤隐藏**：不再显示"正在执行操作：read_file、run_shell"等技术细节
-- **思考与回复分离**：思考过程实时显示，最终回复简洁明了，不重复
-
-### 2. Electron 桌面版正式发布
-
-- **独立桌面窗口**：不再依赖浏览器，双击即可使用
-- **系统托盘支持**：最小化到后台，随时唤起
-- **防止多开**：单实例运行，避免资源浪费
-- **应用菜单**：完整的编辑菜单（复制/粘贴/剪切/全选）
-- **权限管理**：麦克风、摄像头、屏幕捕获权限自动处理
-- **一键启动脚本**：`一键启动桌面版.bat`，自动 build + 启动
-
-### 3. 界面与交互优化
-
-- **界面宽度优化**：主区域最大宽度 1400px，不再占满整个屏幕
-- **用户菜单重构**：从底部上拉改为右侧侧边栏（280px）
-- **启动逻辑修复**：不再打开两次网页，一次启动直接进入
-- **文本选择复制**：支持选中部分文字复制，不再只能整条复制
-- **消息操作按钮**：悬停显示复制、创建文档按钮
-- **对话历史扩容**：conversation 上限从 80 条提升到 300 条，steps 上限从 200 提升到 500 条
-
-### 4. 功能新增与完善
-
-- **截图功能**：支持屏幕捕获 + 框选裁剪，截图直接到输入框
-- **电脑操作**：后端 PowerShell API，支持截图、鼠标移动/点击、键盘输入
-- **语音输入**：Web Speech API，建议 Edge 浏览器使用
-- **语音/视频通话**：麦克风实时识别 + 摄像头画面 + 视频截图
-- **工作区选择器**：驱动器和文件夹分页，支持新建/重命名文件夹
-- **对话任务优化**：点击对话直接新建任务，无需手动点新建
-- **自动化快捷指令**：常用快捷指令内置
-- **模板库插件系统**：支持自定义来源，节点连接外部插件
-- **记忆系统**：长期记忆自定义，添加记录简化为纯文本输入
-
-### 5. 编码与修复能力提升
-
-- **系统提示词升级**：新增编码能力 8 条要求、修复 bug 能力 7 条要求
-- **错误分类优化**：build-error 等错误提供 5 步具体排查流程
-- **连续错误处理**：遍历所有连续错误分类记录，不再只记录最后一个
-- **自检维护优化**：每天仅第一次任务可做环境确认，后续直接执行，出错才修复
-- **构建错误自愈**：失败时给出结构化报告（错误类型+最后错误+修复建议+下一步）
-
-### 6. 中英文国际化
-
-- **后端 i18n**：完整的中英文词条
-- **前端国际化**：电脑操作、视频通话、消息操作等新增中英文翻译
-- **持续完善中**：部分硬编码文本逐步迁移到 i18n
-
-### 7. 本地语音识别服务（可选）
-
-- **faster-whisper 本地部署**：完全免费、离线、无调用限制
-- **HTTP API 接口**：任何项目都能调用
-- **支持多语言**：中文优化，准确率高
-- **独立服务**：不与主程序耦合，按需启动
-
----
-
-## Bug 修复
-
-- 修复 Electron 桌面版自动退出问题
-- 修复启动后打开两次网页的问题
-- 修复对话历史过长导致前面内容被截断的问题
-- 修复任务执行中思考过程不实时显示的问题
-- 修复桌面版复制粘贴快捷键失效的问题
-- 修复桌面版截图权限反复请求的问题
-- 修复 13389.js 的 s.nodeName.toLowerCase 错误（非项目代码，已确认不影响功能）
-- 修复构建错误连续 3 次后错误信息不准确的问题
-
----
-
-## 升级指南
-
-### 从 0.6.1 升级
-
-```bash
-# 拉取最新代码
-git pull origin master
-
-# 安装依赖（如有新增）
-npm install
-
-# 重新构建
-npm run build
-
-# 启动服务
-fhcode serve
-# 或
-node dist/cli/index.js serve
-```
-
-### 桌面版使用
-
-```bash
-# 方式一：一键启动
-双击 一键启动桌面版.bat
-
-# 方式二：手动启动
-npm run build
-npm run electron
-```
-
-### 本地语音识别服务（可选）
-
-```bash
-cd funasr-server
-双击 启动语音识别服务.bat
-# 服务启动在 http://localhost:8082
-```
-
----
-
-## 系统要求
-
-- **Node.js**：18.0 或更高版本
-- **操作系统**：Windows 10/11（推荐）、macOS、Linux
-- **内存**：至少 4GB（推荐 8GB）
-- **桌面版**：Windows 10/11，Electron v41+
-
----
-
-## 已知限制
-
-- 桌面版语音输入依赖 Google 服务，国内建议使用网页版（Edge 浏览器）或本地语音识别服务
-- 电脑操作功能目前为手动操作界面，AI 自动操作正在开发中
-- 中英文国际化尚未完全覆盖所有硬编码文本
-- 本地语音识别服务为非实时识别（录音完成后识别）
-
----
-
-## 反馈与支持
-
-- GitHub Issues：https://github.com/wch887292/feihong-code/issues
-- npm：https://www.npmjs.com/package/feihong-code
-
----
-
-*晋江市飞虹智科技企业管理有限公司 · 飞扬企源研发中心*
-*负责人：吴赐虹*
+# 飞虹 Code v7.0.0 更新说明书
+
+**发布日期**：2026年8月24日
+**版本号**：7.0.0
+**产品名称**：飞虹 Code（feihong-code）
+**研发团队**：晋江市飞虹智科技企业管理有限公司 · 飞扬企源研发中心
+
+---
+
+## 版本概述
+
+飞虹 Code v7.0.0 是一个重大版本升级，从 0.6.1 直接跨越到 7.0.0，标志着产品从终端工具向全平台 AI 编程助手的成熟蜕变。本版本重点优化了用户体验、新增了桌面端支持、完善了多语言国际化，并大幅提升了对话交互的自然度。
+
+---
+
+## 核心亮点
+
+### 1. 对话流全面重构（对标豆包体验）
+
+- **纯文本输出**：彻底去除 Markdown 格式、工具调用状态、技术摘要等干扰信息
+- **实时思考过程**：模型思考过程实时展示，让用户看到 AI 在想什么
+- **口语化表达**：像普通人交流一样自然，不再是技术报告风格
+- **操作步骤隐藏**：不再显示"正在执行操作：read_file、run_shell"等技术细节
+- **思考与回复分离**：思考过程实时显示，最终回复简洁明了，不重复
+
+### 2. Electron 桌面版正式发布
+
+- **独立桌面窗口**：不再依赖浏览器，双击即可使用
+- **系统托盘支持**：最小化到后台，随时唤起
+- **防止多开**：单实例运行，避免资源浪费
+- **应用菜单**：完整的编辑菜单（复制/粘贴/剪切/全选）
+- **权限管理**：麦克风、摄像头、屏幕捕获权限自动处理
+- **一键启动脚本**：`一键启动桌面版.bat`，自动 build + 启动
+
+### 3. 界面与交互优化
+
+- **界面宽度优化**：主区域最大宽度 1400px，不再占满整个屏幕
+- **用户菜单重构**：从底部上拉改为右侧侧边栏（280px）
+- **启动逻辑修复**：不再打开两次网页，一次启动直接进入
+- **文本选择复制**：支持选中部分文字复制，不再只能整条复制
+- **消息操作按钮**：悬停显示复制、创建文档按钮
+- **对话历史扩容**：conversation 上限从 80 条提升到 300 条，steps 上限从 200 提升到 500 条
+
+### 4. 功能新增与完善
+
+- **截图功能**：支持屏幕捕获 + 框选裁剪，截图直接到输入框
+- **电脑操作**：后端 PowerShell API，支持截图、鼠标移动/点击、键盘输入
+- **语音输入**：Web Speech API，建议 Edge 浏览器使用
+- **语音/视频通话**：麦克风实时识别 + 摄像头画面 + 视频截图
+- **工作区选择器**：驱动器和文件夹分页，支持新建/重命名文件夹
+- **对话任务优化**：点击对话直接新建任务，无需手动点新建
+- **自动化快捷指令**：常用快捷指令内置
+- **模板库插件系统**：支持自定义来源，节点连接外部插件
+- **记忆系统**：长期记忆自定义，添加记录简化为纯文本输入
+
+### 5. 编码与修复能力提升
+
+- **系统提示词升级**：新增编码能力 8 条要求、修复 bug 能力 7 条要求
+- **错误分类优化**：build-error 等错误提供 5 步具体排查流程
+- **连续错误处理**：遍历所有连续错误分类记录，不再只记录最后一个
+- **自检维护优化**：每天仅第一次任务可做环境确认，后续直接执行，出错才修复
+- **构建错误自愈**：失败时给出结构化报告（错误类型+最后错误+修复建议+下一步）
+
+### 6. 中英文国际化
+
+- **后端 i18n**：完整的中英文词条
+- **前端国际化**：电脑操作、视频通话、消息操作等新增中英文翻译
+- **持续完善中**：部分硬编码文本逐步迁移到 i18n
+
+### 7. 本地语音识别服务（可选）
+
+- **faster-whisper 本地部署**：完全免费、离线、无调用限制
+- **HTTP API 接口**：任何项目都能调用
+- **支持多语言**：中文优化，准确率高
+- **独立服务**：不与主程序耦合，按需启动
+
+---
+
+## Bug 修复
+
+- 修复 Electron 桌面版自动退出问题
+- 修复启动后打开两次网页的问题
+- 修复对话历史过长导致前面内容被截断的问题
+- 修复任务执行中思考过程不实时显示的问题
+- 修复桌面版复制粘贴快捷键失效的问题
+- 修复桌面版截图权限反复请求的问题
+- 修复 13389.js 的 s.nodeName.toLowerCase 错误（非项目代码，已确认不影响功能）
+- 修复构建错误连续 3 次后错误信息不准确的问题
+
+---
+
+## 升级指南
+
+### 从 0.6.1 升级
+
+```bash
+# 拉取最新代码
+git pull origin master
+
+# 安装依赖（如有新增）
+npm install
+
+# 重新构建
+npm run build
+
+# 启动服务
+fhcode serve
+# 或
+node dist/cli/index.js serve
+```
+
+### 桌面版使用
+
+```bash
+# 方式一：一键启动
+双击 一键启动桌面版.bat
+
+# 方式二：手动启动
+npm run build
+npm run electron
+```
+
+### 本地语音识别服务（可选）
+
+```bash
+cd funasr-server
+双击 启动语音识别服务.bat
+# 服务启动在 http://localhost:8082
+```
+
+---
+
+## 系统要求
+
+- **Node.js**：18.0 或更高版本
+- **操作系统**：Windows 10/11（推荐）、macOS、Linux
+- **内存**：至少 4GB（推荐 8GB）
+- **桌面版**：Windows 10/11，Electron v41+
+
+---
+
+## 已知限制
+
+- 桌面版语音输入依赖 Google 服务，国内建议使用网页版（Edge 浏览器）或本地语音识别服务
+- 电脑操作功能目前为手动操作界面，AI 自动操作正在开发中
+- 中英文国际化尚未完全覆盖所有硬编码文本
+- 本地语音识别服务为非实时识别（录音完成后识别）
+
+---
+
+## 反馈与支持
+
+- GitHub Issues：https://github.com/wch887292/feihong-code/issues
+- npm：https://www.npmjs.com/package/feihong-code
+
+---
+
+*晋江市飞虹智科技企业管理有限公司 · 飞扬企源研发中心*
+*负责人：吴赐虹*