@geminilight/mindos 0.1.7 → 0.1.9

package/README.md

@@ -65,42 +65,45 @@ Static documents are hard to synchronize and weak as execution systems in real h
 
 ## ✨ Features
 
-### For Humans
+**For Humans**
 
 - **GUI Collaboration Workbench**: use one command entry to browse, edit, and search efficiently (`⌘K` / `⌘/`).
 - **Built-in Agent Assistant**: converse in context while edits are captured into managed knowledge.
 - **Plugin Views**: use scenario-focused views like TODO, Kanban, and Timeline.
 
-### For Agents
+**For Agents**
 
 - **MCP Server + Skills**: connect any compatible agent to read, write, search, and run workflows.
 - **Structured Templates**: start quickly with Profile, Workflows, and Configurations scaffolds.
 - **Experience Auto-Distillation**: automatically distill daily work into reusable, executable SOP experience.
 
-### Infrastructure
+**Infrastructure**
 
 - **Reference Sync**: keep cross-file status and context aligned via links/backlinks.
 - **Knowledge Graph**: visualize relationships and dependencies across notes.
 - **Git Time Machine**: track every edit, audit history, and roll back safely.
 
-**Coming Soon:**
+<details>
+<summary><strong>Coming Soon</strong></summary>
 
 - [ ] ACP (Agent Communication Protocol): connect external Agents (e.g., Claude Code, Cursor) and turn the knowledge base into a multi-Agent collaboration hub
 - [ ] Deep RAG integration: retrieval-augmented generation grounded in your knowledge base for more accurate, context-aware AI responses
 - [ ] Backlinks View: display all files that reference the current file, helping you understand how a note fits into the knowledge network
 - [ ] Agent Inspector: render Agent operation logs as a filterable timeline to audit every tool call in detail
 
+</details>
+
 ---
 
 ## 🚀 Getting Started
 
 > [!IMPORTANT]
-> **Quick Start with Agent:** Paste this prompt into any MCP-capable Agent (Claude Code, Cursor, etc.) to install automatically, then skip to [Step 4](#4-inject-your-personal-mind-with-mindos-agent):
+> **Quick Start with Agent:** Paste this prompt into any MCP-capable Agent (Claude Code, Cursor, etc.) to install automatically, then skip to [Step 3](#3-inject-your-personal-mind-with-mindos-agent):
 > ```
 > Help me install MindOS from https://github.com/GeminiLight/MindOS with MCP and Skills. Use English template.
 > ```
 
-> Already have a knowledge base? Skip to [Step 5](#5-make-any-agent-ready-mcp--skills) to configure MCP + Skills.
+> Already have a knowledge base? Skip to [Step 4](#4-make-any-agent-ready-mcp--skills) to configure MCP + Skills.
 
 ### 1. Install
 
@@ -129,10 +132,13 @@ mindos onboard --install-daemon
 
 The setup wizard will guide you through:
 1. Knowledge base path → default `~/.mindos/my-mind`
-2. Choose template language (en / zh)
-3. Optional Auth token to protect your API
-4. Configure AI Provider (Anthropic / OpenAI) + API Key — or choose **skip** to configure later via `mindos config set`
-5. Config saved to `~/.mindos/config.json` automatically
+2. Choose template (en / zh / empty / custom)
+3. Ports (Web UI + MCP)
+4. Auth token (auto-generated or passphrase-seeded)
+5. Web UI password (optional)
+6. AI Provider (Anthropic / OpenAI) + API Key — or **skip** to configure later via `mindos config set`
+
+Config is saved to `~/.mindos/config.json` automatically.
 
 <!-- For Agents: Before running setup, if the user has not provided LLM configuration, ask them:
 1. AI Provider — `anthropic` or `openai` (compatible)?
@@ -183,24 +189,10 @@ Multiple providers can be configured simultaneously — switch between them by c
 > [!NOTE]
 > If you want the MindOS GUI to be reachable from other devices, make sure the port is open in firewall/security-group settings and bound to an accessible host/network interface.
 
-### 3. Run
-
-```bash
-mindos start    # start app + MCP server (production, auto-rebuilds when a new version is detected)
-```
-
-On startup, MindOS prints your Web UI URL, MCP config snippet, and auth token — ready to copy into your Agent.
-
 > [!TIP]
-> **Upgrading?** Run `mindos update` to install the latest version.
+> With `--install-daemon`, MindOS is installed as a background OS service and starts automatically — no need to run `mindos start` manually. If you skipped the flag, run `mindos start` to start manually, or `mindos update` to upgrade to the latest version.
 
-> [!TIP]
-> If the port is already in use, set a different port during `mindos onboard` or update `~/.mindos/config.json`:
-> ```bash
-> MINDOS_WEB_PORT=3002 mindos
-> ```
-
-### 4. Inject Your Personal Mind with MindOS Agent
+### 3. Inject Your Personal Mind with MindOS Agent
 
 1. Open the built-in MindOS Agent chat panel in the GUI.
 2. Upload your resume or any personal/project material.
@@ -210,24 +202,51 @@ On startup, MindOS prints your Web UI URL, MCP config snippet, and auth token
   <img src="assets/images/gui-sync-cv.png" alt="Sync CV Example" width="800" />
 </p>
 
-### 5. Make Any Agent Ready (MCP + Skills)
+### 4. Make Any Agent Ready (MCP + Skills)
 
-#### 5.1 Configure MindOS MCP
+#### 4.1 Configure MindOS MCP
 
 <!-- For Agents: Register MindOS MCP as global (user-level) config by default, so all projects share the same knowledge base. -->
 
-> [!TIP]
-> Each Agent stores MCP config in a different file — see the **MCP Config Path** column in the [Supported Agents](#-supported-agents) table below for exact paths.
+Run one command to auto-install MCP config into your Agent:
 
-Register the MindOS MCP Server in your Agent client:
+```bash
+mindos mcp install
+```
+
+Supports Claude Code, Cursor, Windsurf, Cline, Trae, Gemini CLI, and more — interactively guides you through agent, scope, transport, and token.
 
-The MCP server starts automatically when you run `mindos start` or `mindos dev`. The MCP URL, auth token, and a ready-to-paste config snippet are printed on every startup.
+**Local (Agent and MindOS on the same machine)**
 
-> The MCP port defaults to `8787`. To use a different port, set `mcpPort` in `~/.mindos/config.json` via `mindos onboard`.
+Use `stdio` transport — no server process needed, most reliable:
 
-**Option A: Local (same machine)**
+```bash
+# Interactive
+mindos mcp install
 
-Via stdio — no server process needed:
+# One-shot, global scope (shared across all projects)
+mindos mcp install -g -y
+```
+
+**Remote (Agent on a different machine)**
+
+Use `http` transport — MindOS must be running (`mindos start`) on the remote machine:
+
+```bash
+mindos mcp install--transport http --url http://<server-ip>:8787/mcp --token your-token -g
+```
+
+> [!NOTE]
+> For remote access, ensure port `8787` is open in your firewall/security-group.
+
+> Add `-g` to install globally — MCP config is shared across all projects instead of the current directory only.
+
+> The MCP port defaults to `8787`. To use a different port, run `mindos onboard` and set `mcpPort`.
+
+<details>
+<summary>Manual config (JSON snippets)</summary>
+
+**Local via stdio** (no server process needed):
 
 ```json
 {
@@ -242,7 +261,7 @@ Via stdio — no server process needed:
 }
 ```
 
-Or via URL:
+**Local via URL:**
 
 ```json
 {
@@ -255,10 +274,7 @@ Or via URL:
 }
 ```
 
-**Option B: Remote URL (another device)**
-
-> [!NOTE]
-> Ensure port `8787` is open in your firewall/security-group so remote clients can reach the server.
+**Remote:**
 
 ```json
 {
@@ -271,7 +287,11 @@ Or via URL:
 }
 ```
 
-#### 5.2 Install MindOS Skills
+Each Agent stores config in a different file — see the **MCP Config Path** column in the [Supported Agents](#-supported-agents) table for exact paths.
+
+</details>
+
+#### 4.2 Install MindOS Skills
 
 | Skill | Description |
 |-------|-------------|
@@ -290,7 +310,7 @@ npx skills add https://github.com/GeminiLight/MindOS --skill mindos-zh -g -y
 
 MCP = connection capability, Skills = workflow capability. Enabling both gives the complete MindOS agent experience.
 
-#### 5.3 Common Pitfalls
+#### 4.3 Common Pitfalls
 
 - Only MCP, no Skills: tools are callable, but best-practice workflows are missing.
 - Only Skills, no MCP: workflow guidance exists, but the Agent cannot operate your local knowledge base.

package/README_zh.md

@@ -65,42 +65,45 @@ MindOS 是一个**人机协同心智系统**——基于本地优先的协作知
 
 ## ✨ 功能特性
 
-### 人类侧
+**人类侧**
 
 - **GUI 协作工作台**：以统一入口高效浏览、编辑与搜索（`⌘K` / `⌘/`）。
 - **内置 Agent 助手**：在上下文中对话，编辑内容可持续沉淀为可管理知识。
 - **插件视图**：按场景使用 TODO、看板、时间线等视图。
 
-### Agent 侧
+**Agent 侧**
 
 - **MCP Server + Skills**：让兼容 Agent 统一接入读写、搜索与工作流执行。
 - **结构化模板**：通过 Profile、Workflows、Configurations 快速冷启动。
 - **经验自动沉淀**：将日常记录自动化沉淀为可执行 SOP 经验。
 
-### 基础设施
+**基础设施**
 
 - **引用同步**：通过引用与反向链接保持跨文件状态一致。
 - **知识图谱**：可视化笔记间关系与依赖。
 - **Git 时光机**：记录修改历史，支持审计与安全回滚。
 
-**即将到来：**
+<details>
+<summary><strong>即将到来</strong></summary>
 
 - [ ] ACP（Agent Communication Protocol）：连接外部 Agent（如 Claude Code、Cursor），让知识库成为多 Agent 协作的中枢
 - [ ] RAG 深度集成：基于知识库内容的检索增强生成，让 AI 回答更精准、更有上下文
 - [ ] 反向链接视图（Backlinks）：展示所有引用当前文件的反向链接，理解笔记在知识网络中的位置
 - [ ] Agent 审计面板（Agent Inspector）：将 Agent 操作日志渲染为可筛选的时间线，审查每次工具调用的详情
 
+</details>
+
 ---
 
 ## 🚀 快速开始
 
 > [!IMPORTANT]
-> **用 Agent 一键安装：** 将以下指令粘贴到任意支持 MCP 的 Agent（Claude Code、Cursor 等），即可自动完成安装，然后跳到[第 4 步](#4-通过-mindos-agent-注入你的个人心智)：
+> **用 Agent 一键安装：** 将以下指令粘贴到任意支持 MCP 的 Agent（Claude Code、Cursor 等），即可自动完成安装，然后跳到[第 3 步](#3-通过-mindos-agent-注入你的个人心智)：
 > ```
 > 帮我从 https://github.com/GeminiLight/MindOS 安装 MindOS，包含 MCP 和 Skills，使用中文模板。
 > ```
 
-> 已有知识库？直接跳到[第 5 步](#5-让任意-agent-可用mcp--skills)配置 MCP + Skills。
+> 已有知识库？直接跳到[第 4 步](#4-让任意-agent-可用mcp--skills)配置 MCP + Skills。
 
 ### 1. 安装
 
@@ -129,10 +132,13 @@ mindos onboard --install-daemon
 
 配置向导将引导你完成：
 1. 知识库路径 → 默认 `~/.mindos/my-mind`
-2. 选择模板语言（en / zh / empty）
-3. 可选 Auth token，保护你的 API
-4. 配置 AI Provider（Anthropic / OpenAI）+ API Key — 或选择 **skip**，稍后通过 `mindos config set` 补填
-5. 配置自动保存到 `~/.mindos/config.json`
+2. 选择模板（en / zh / empty / custom）
+3. 端口配置（Web UI + MCP）
+4. Auth token（自动生成或口令派生）
+5. Web UI 访问密码（可选）
+6. 配置 AI Provider（Anthropic / OpenAI）+ API Key — 或选择 **skip**，稍后通过 `mindos config set` 补填
+
+配置自动保存到 `~/.mindos/config.json`。
 
 <!-- For Agents: 运行 onboard 前，如果用户未提供 LLM 配置，请询问：
 1. AI Provider — `anthropic` 还是 `openai`（兼容）？
@@ -183,24 +189,10 @@ mindos onboard --install-daemon
 > [!NOTE]
 > 如果你希望其他设备也能访问 MindOS GUI，请确保端口已在防火墙/安全组中放行，并绑定到可访问的主机地址/网卡。
 
-### 3. 启动
-
-```bash
-mindos start    # 启动 app + MCP 服务（生产模式，检测到新版本时自动重新构建）
-```
-
-启动后，MindOS 会自动打印 Web UI 地址、MCP 配置片段和 Auth token，直接复制到你的 Agent 即可。
-
 > [!TIP]
-> **升级？** 运行 `mindos update` 安装最新版本。
+> 使用 `--install-daemon` 时，MindOS 会作为后台 OS 服务安装并自动启动，无需手动执行 `mindos start`。如果跳过了该参数，运行 `mindos start` 手动启动，或运行 `mindos update` 升级到最新版本。
 
-> [!TIP]
-> 如果端口被占用，在 `mindos onboard` 时指定其他端口，或更新 `~/.mindos/config.json`，也可通过环境变量临时覆盖：
-> ```bash
-> MINDOS_WEB_PORT=3002 mindos
-> ```
-
-### 4. 通过 MindOS Agent 注入你的个人心智
+### 3. 通过 MindOS Agent 注入你的个人心智
 
 1. 打开 MindOS GUI 中内置的 Agent 对话面板。
 2. 上传你的简历或任意个人/项目资料。
@@ -210,24 +202,51 @@ mindos start    # 启动 app + MCP 服务（生产模式，检测到新版本时
   <img src="assets/images/gui-sync-cv.png" alt="同步简历示例" width="800" />
 </p>
 
-### 5. 让任意 Agent 可用（MCP + Skills）
+### 4. 让任意 Agent 可用（MCP + Skills）
 
-#### 5.1 配置 MindOS MCP
+#### 4.1 配置 MindOS MCP
 
 <!-- For Agents: MindOS MCP 默认注册为全局（user-level）配置，使所有项目共享同一知识库。-->
 
-> [!TIP]
-> 每个 Agent 的 MCP 配置文件路径不同——详见下方 [支持的 Agent](#-支持的-agent) 表格中的 **MCP 配置文件路径** 列。
+一条命令自动将 MCP 配置写入 Agent：
 
-将 MindOS MCP Server 注册到你的 Agent 客户端：
+```bash
+mindos mcp install
+```
+
+支持 Claude Code、Cursor、Windsurf、Cline、Trae、Gemini CLI 等主流 Agent，交互式引导完成 agent、scope、transport 和 token 的配置。
 
-运行 `mindos start` 或 `mindos dev` 时，MCP 服务自动随之启动。每次启动都会打印 MCP URL、Auth token 及可直接粘贴的配置片段。
+**本机访问（Agent 和 MindOS 在同一台机器）**
 
-> MCP 端口默认为 `8787`。如需修改，通过 `mindos onboard` 设置 `~/.mindos/config.json` 中的 `mcpPort`。
+使用 `stdio` transport — 无需启动服务进程，最稳定：
 
-**方式 A：本机（同一台机器）**
+```bash
+# 交互式
+mindos mcp install
 
-通过 stdio — 无需启动服务进程：
+# 一键安装，全局（所有项目共享）
+mindos mcp install -g -y
+```
+
+**远程访问（Agent 在另一台机器）**
+
+使用 `http` transport — 远程机器上需先运行 `mindos start`：
+
+```bash
+mindos mcp install --transport http --url http://<服务器IP>:8787/mcp --token your-token -g
+```
+
+> [!NOTE]
+> 远程访问时，请确保端口 `8787` 已在防火墙/安全组中放行。
+
+> 加 `-g` 表示全局安装 — MCP 配置写入用户级配置文件，所有项目共享，而非仅当前目录。
+
+> MCP 端口默认为 `8787`。如需修改，运行 `mindos onboard` 设置 `mcpPort`。
+
+<details>
+<summary>手动配置（JSON 片段）</summary>
+
+**本机 stdio**（无需启动服务进程）：
 
 ```json
 {
@@ -242,7 +261,7 @@ mindos start    # 启动 app + MCP 服务（生产模式，检测到新版本时
 }
 ```
 
-或通过 URL：
+**本机 URL：**
 
 ```json
 {
@@ -255,10 +274,7 @@ mindos start    # 启动 app + MCP 服务（生产模式，检测到新版本时
 }
 ```
 
-**方式 B：远程 URL（其他设备）**
-
-> [!NOTE]
-> 请确保端口 `8787` 已在防火墙/安全组中放行，远程客户端才能连接。
+**远程：**
 
 ```json
 {
@@ -271,7 +287,11 @@ mindos start    # 启动 app + MCP 服务（生产模式，检测到新版本时
 }
 ```
 
-#### 5.2 安装 MindOS Skills
+各 Agent 的配置文件路径不同，详见下方 [支持的 Agent](#-支持的-agent) 表格中的 **MCP 配置文件路径** 列。
+
+</details>
+
+#### 4.2 安装 MindOS Skills
 
 | Skill | 说明 |
 |-------|------|
@@ -290,7 +310,7 @@ npx skills add https://github.com/GeminiLight/MindOS --skill mindos-zh -g -y
 
 MCP = 连接能力，Skills = 工作流能力；两者都开启后体验完整。
 
-#### 5.3 常见误区
+#### 4.3 常见误区
 
 - 只配 MCP，不装 Skills：能调用工具，但缺少最佳实践指引。
 - 只装 Skills，不配 MCP：有流程提示，但无法操作本地知识库。

package/assets/demo-flow-zh.html

@@ -358,8 +358,8 @@
 <body>
 <div class="canvas" id="canvas">
   <div class="header">
-    <h1>MindOS：从灵感到执行到复盘</h1>
-    <p>随手记录 &rarr; 自动整理 &rarr; 所有 Agent 同步 &rarr; 经验回流</p>
+    <h1>MindOS：想法、执行到复盘，一条线贯穿</h1>
+    <p>记一句想法，剩下的全自动</p>
   </div>
 
   <div class="flow" id="flow">
@@ -375,18 +375,18 @@
         <div class="mobile-nav"><span class="back">&lsaquo;</span><span class="title">Quick Capture</span></div>
         <div class="mobile-body">
           <div class="mobile-note">
-            <div class="note-time">今天 09:41</div>
-            <div class="note-text">用户反馈：新手引导太长，50% 的人<br/>在第 3 步就放弃了。要简化到 2 步内<br/>完成首次体验，参考 Linear 的做法。</div>
-            <div class="note-tag">#用户反馈</div>
+            <div class="note-time">今天 11:30</div>
+            <div class="note-text">想好了新项目：先理清思路，<br/>代码开搞，顺便发个帖宣传下</div>
+            <div class="note-tag">#想法</div>
           </div>
           <div class="mobile-note" style="border-left-color: var(--muted-fg);">
-            <div class="note-time">昨天 23:20</div>
-            <div class="note-text">灵感：把安装流程做成"一句话发给<br/>Agent"，零摩擦。方便用户直接上手。</div>
-            <div class="note-tag" style="background:rgba(138,130,117,0.1);color:var(--muted-fg);">#产品灵感</div>
+            <div class="note-time">昨天 16:00</div>
+            <div class="note-text">之前调研过竞品，用 React + Tailwind 方案</div>
+            <div class="note-tag" style="background:rgba(138,130,117,0.1);color:var(--muted-fg);">#技术选型</div>
           </div>
         </div>
         <div class="mobile-input">
-          <input type="text" value="写个用户增长方案..." readonly />
+          <input type="text" value="记一下新项目的想法..." readonly />
           <div class="send-btn">&uarr;</div>
         </div>
       </div>
@@ -410,30 +410,30 @@
         <div class="desktop-layout">
           <div class="sidebar">
             <div class="sidebar-item"><span class="icon">P</span> Profile</div>
-            <div class="sidebar-item active"><span class="icon">W</span> Workflows</div>
-            <div class="sidebar-item"><span class="icon">J</span> Projects</div>
+            <div class="sidebar-item"><span class="icon">W</span> Workflows</div>
+            <div class="sidebar-item active"><span class="icon">J</span> Projects</div>
             <div class="sidebar-item"><span class="icon">C</span> Configs</div>
             <div class="sidebar-item"><span class="icon">R</span> Resources</div>
             <div class="sidebar-item"><span class="icon">G</span> Graph</div>
           </div>
           <div class="main-content">
-            <div class="file-path">Workflows/用户增长策略.md</div>
+            <div class="file-path">Projects/新项目计划.md</div>
             <div class="updated-tag">* Agent 已自动更新 3 个文件</div>
-            <div class="content-title">用户增长策略</div>
+            <div class="content-title">新项目计划</div>
             <div class="content-section">
-              <h3>核心行动项</h3>
+              <h3>核心思路</h3>
               <ul>
-                <li>简化新手引导：3 步压缩为 1 步</li>
-                <li>一句话安装作为核心传播点</li>
-                <li>制作 Before/After 对比视频</li>
+                <li>先理清项目方案和架构</li>
+                <li>搭建代码骨架，快速出原型</li>
+                <li>发帖宣传，收集早期反馈</li>
               </ul>
             </div>
             <div class="content-section">
               <h3>关联更新</h3>
               <div class="badge-row">
-                <span class="badge">Projects/增长实验.md</span>
-                <span class="badge">Profile/产品理念.md</span>
-                <span class="badge">Workflows/发布SOP.md</span>
+                <span class="badge">Projects/Project-Plan.md</span>
+                <span class="badge">Workflows/Launch-SOP.md</span>
+                <span class="badge">TODO.md</span>
               </div>
             </div>
           </div>
@@ -458,12 +458,12 @@
           <div class="cli-frame">
             <div class="cli-titlebar">
               <div class="dots"><span class="dot-r-sm"></span><span class="dot-y-sm"></span><span class="dot-g-sm"></span></div>
-              <span class="name">Claude Code</span>
+              <span class="name">Gemini</span>
             </div>
             <div class="cli-body">
-              <div class="cli-line"><span class="cli-prompt">&gt; </span><span class="cli-cmd">写一篇产品推广文案</span></div>
-              <div class="cli-line"><span class="cli-reading">  Reading MindOS &rarr; 用户增长策略.md</span></div>
-              <div class="cli-line"><span class="cli-output">  Done. 已基于增长策略生成推广文案</span></div>
+              <div class="cli-line"><span class="cli-prompt">&gt; </span><span class="cli-cmd">梳理项目方案和架构</span></div>
+              <div class="cli-line"><span class="cli-reading">  Reading MindOS &rarr; Project-Plan.md</span></div>
+              <div class="cli-line"><span class="cli-output">  Done. 已整理项目方案 + 技术架构文档</span></div>
             </div>
           </div>
           <!-- Cursor -->
@@ -473,9 +473,9 @@
               <span class="name">Cursor</span>
             </div>
             <div class="cli-body">
-              <div class="cli-line"><span class="cli-prompt">&gt; </span><span class="cli-cmd">简化新手引导到 1 步完成</span></div>
-              <div class="cli-line"><span class="cli-reading">  Reading MindOS &rarr; 产品理念.md</span></div>
-              <div class="cli-line"><span class="cli-output">  Done. 已重构 onboarding 组件</span></div>
+              <div class="cli-line"><span class="cli-prompt">&gt; </span><span class="cli-cmd">搭建项目骨架</span></div>
+              <div class="cli-line"><span class="cli-reading">  Reading MindOS &rarr; Project-Plan.md</span></div>
+              <div class="cli-line"><span class="cli-output">  Done. 已初始化项目结构 + 基础配置</span></div>
             </div>
           </div>
           <!-- OpenClaw (Telegram style) -->
@@ -486,8 +486,8 @@
               <span class="tg-badge">BOT</span>
             </div>
             <div class="tg-body">
-              <div class="tg-bubble user">帮我把这轮增长实验的经验沉淀下来</div>
-              <div class="tg-bubble bot"><span class="tg-hl">Writing &rarr; Workflows/增长复盘.md</span><br/>Done. 已沉淀为可复用 SOP</div>
+              <div class="tg-bubble user">发帖宣传项目，收集大家反馈</div>
+              <div class="tg-bubble bot"><span class="tg-hl">Writing &rarr; Workflows/Launch-SOP.md</span><br/>Done. 已发布宣传帖 + 汇总反馈沉淀到 SOP</div>
             </div>
           </div>
         </div>
@@ -498,7 +498,7 @@
   <svg class="arrows-svg" id="arrows-svg" xmlns="http://www.w3.org/2000/svg"></svg>
 
   <div class="tagline">
-    <p>你 &rarr; <span>MindOS</span> &rarr; 所有 Agent &rarr; <span>经验回流</span> &rarr; 持续进化的共享心智</p>
+    <p>一句想法 &rarr; <span>MindOS</span> 归档 + 关联 &rarr; 所有 Agent 各就各位 &rarr; <span>经验自动沉淀</span></p>
   </div>
 
   <div class="cta-card">