@geminilight/mindos-linux-arm64-musl 1.1.58 → 1.1.60

package/README.md

@@ -24,7 +24,7 @@
   <a href="https://github.com/GeminiLight/MindOS/releases?q=clipper"><img src="https://img.shields.io/badge/Web_Clipper-Install-c8873a.svg?style=for-the-badge" alt="Web Clipper"></a>
 </p>
 
-MindOS is where you think, and where your AI agents act — a local-first knowledge base shared between you and every AI you use. **Share your brain with every AI — every thought grows.**
+MindOS is a local-first knowledge base for sharing durable context between you and the AI agents you use. **Think in one place. Let agents work from shared context.**
 
 ---
 
@@ -78,7 +78,7 @@ MindOS is where you think, and where your AI agents act — a local-first knowle
 </table>
 
 > [!IMPORTANT]
-> **⭐ One-click install:** Send this to your Agent (Claude Code, Cursor, etc.) to set up everything automatically:
+> **⭐ Agent-assisted install:** Send this to your Agent (Claude Code, Cursor, etc.) to set up MindOS, MCP, and Skills:
 > ```
 > Help me install MindOS from https://github.com/GeminiLight/MindOS with MCP and Skills. Use English template.
 > ```
@@ -96,14 +96,15 @@ MindOS is where you think, and where your AI agents act — a local-first knowle
 
 ## 📢 What's New
 
-| Date | Update |
-|------|--------|
-| 2026-04 | **Protocol runtimes internalized** — ACP and MCP source now live under `packages/mindos/src/protocols/*`; the published package ships self-contained `dist/protocols/*` bundles instead of nested protocol packages. |
-| 2026-04 | **Product main package** — MindOS now follows an OpenCode-style core package: `packages/mindos` (`@geminilight/mindos`) is the product runtime facade and CLI kernel boundary; Web/CLI adapt around it. |
-| 2026-04 | **OpenCode-style workspace flattening** — all source workspaces now live under `packages/`: `packages/web`, `packages/desktop`, `packages/mobile`, `packages/mindos`, and optional `packages/retrieval/*` adapters. npm tarball, CLI, MCP, Web standalone, Desktop, and Mobile checks have passed. |
-| 2026-04 | **Clean publish package** — legacy top-level `app/`, `mcp/`, `desktop/`, and `mobile/` source roots are removed; publish entries exclude tests, caches, and nested `node_modules`. |
-| 2026-04 | **Web Clipper** — Browser extension to clip any web page to your MindOS Inbox with one click. [Get it →](packages/browser-extension/) |
-| 2026-04 | **PDF & Image Import** — Drag PDFs and images into MindOS, auto-converted to searchable Markdown. |
+MindOS is now organized around a few stable product modules instead of one-off setup scripts:
+
+| Module | What is live |
+|------|------|
+| **Local Knowledge Workbench** | A desktop/web/CLI workspace for browsing, editing, importing, searching, and organizing a local Markdown knowledge base. |
+| **MCP + Skills Bridge** | One command installs or repairs both MCP config and the packaged MindOS Skill, then `mindos doctor agents` verifies real agent readiness. |
+| **Native Agent Runtime** | Run agents with resumable sessions, cancellation, reconnect/reattach support, and reviewable output history where the runtime exposes it. |
+| **Workflow & Audit Surfaces** | YAML workflows, Agent Inspector, run ledgers, backlinks, graph views, and importer progress help keep human review in the loop. |
+| **Cross-platform Delivery** | Install through Desktop releases or `npm install -g @geminilight/mindos@latest`; the npm package includes the CLI and prebuilt local Web runtime. |
 
 ## 🧠 Human-AI Shared Mind
 
@@ -111,24 +112,24 @@ MindOS is where you think, and where your AI agents act — a local-first knowle
 
 **1. Global Sync — Breaking Memory Silos**
 
-Switch tools or start a new chat and you're re-transporting context, scattering knowledge. **With a built-in MCP server, MindOS connects all Agents to your core knowledge base with zero config. Record profile and project memory once to empower all AI tools.**
+Switch tools or start a new chat and you're re-transporting context, scattering knowledge. **With a built-in MCP server and packaged Skills, MindOS gives supported Agents a shared path into your core knowledge base. Record profile and project memory once, then reuse it across your AI tools.**
 
 **2. Transparent & Controllable — No Black Boxes**
 
-Agent memory locked in black boxes makes reasoning unauditable, erasing trust as hallucinations compound. **MindOS saves every retrieval, reflection & action as local plain text. You hold absolute mind-correction rights with a full GUI to recalibrate Agents anytime.**
+Agent memory locked in black boxes makes reasoning hard to audit and correct. **MindOS keeps your knowledge base as local plain text and exposes review surfaces for agent runs, file changes, and important tool activity, so you can inspect and adjust the context agents rely on.**
 
 **3. Symbiotic Evolution — Experience Flows Back As Instructions**
 
-You express preferences but the next chat starts from zero, leaving your thinking useless for AI. **MindOS auto-distills every thought into your knowledge base. Clarify your standards through interaction and sharpen your cognition with each iteration — AI will never repeat the same mistake.**
+You express preferences but the next chat often starts from zero, leaving useful judgment behind. **MindOS helps turn reviewed conversations, corrections, and project standards into reusable notes, Skills, and SOPs, so future agent work can start from a better baseline.**
 
-> **Foundation:** Local-first by default — all data stays in local plain text for privacy, ownership, and speed.
+> **Foundation:** Local-first by default — your knowledge base stays in local plain text for privacy, ownership, and speed.
 
 ---
 
 ## 🚀 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 3](#3-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 MindOS, MCP, and Skills, 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.
 > ```
@@ -186,18 +187,18 @@ mindos open
 
 ### 4. Make Any Agent Ready (MCP + Skills)
 
-**MCP** (connection) — one command to auto-install:
+**MCP + Skills** (connection + workflow) — one command to install or repair both:
 
 ```bash
 mindos mcp install        # interactive
 mindos mcp install -g -y  # one-shot, global scope
 ```
 
-**Skills** (workflow) — install one based on your language:
+Verify the actual agent-facing state:
 
 ```bash
-npx skills add https://github.com/GeminiLight/MindOS --skill mindos -g -y      # English
-npx skills add https://github.com/GeminiLight/MindOS --skill mindos-zh -g -y   # Chinese
+mindos doctor agents
+mindos doctor agents codex --json
 ```
 
 > For remote access, manual JSON config, and common pitfalls, see **[docs/en/supported-agents.md](docs/en/supported-agents.md)**.
@@ -223,28 +224,28 @@ MindOS knowledge doesn't have to start from scratch — bring in what you alread
 
 **For Agents**
 
-- **MCP Server + Skills**: stdio + HTTP dual transport, full-lineup Agent compatible (Claude Code, Cursor, Gemini CLI, etc.). Zero-config access.
-- **ACP / A2A Protocols**: Agent communication protocols for discovery, delegation, and orchestration. Phase 1 is live with Agent Card discovery + JSON-RPC messaging.
-- **Workflow Orchestration**: YAML-based workflow editor with step execution engine — define, edit, and run multi-step agent workflows visually.
+- **MCP Server + Skills**: stdio + HTTP transports for supported Agents (Claude Code, Cursor, Gemini CLI, Codex, and more), with installer/doctor commands for setup and repair.
+- **ACP / A2A Foundations**: early protocol surfaces for agent discovery, registry, session detection, and JSON-RPC messaging. Deeper coordination such as durable mailbox/task-board workflows is still experimental.
+- **Workflow Orchestration**: YAML-based workflow editor with a step execution engine to define, edit, and run repeatable agent workflows.
 - **Structured Templates**: pre-set directory structures for Profiles, Workflows, Configurations, etc., to jumpstart personal context.
-- **Agent-Ready Docs**: everyday notes naturally double as high-quality executable Agent commands — no format conversion needed, write and dispatch.
+- **Agent-Ready Docs**: everyday notes can become reusable Agent instructions — write context once, then dispatch or refine it.
 
 **Infrastructure**
 
 - **Security**: Bearer Token auth, path sandboxing, INSTRUCTION.md write-protection, atomic writes.
 - **Knowledge Graph**: dynamically parses and visualizes inter-file references and dependencies.
 - **Backlinks View**: displays all files that reference the current file, helping you understand how a note fits into the knowledge network.
-- **Agent Inspector**: renders Agent operation logs as a filterable timeline to audit every tool call in detail.
-- **Git Time Machine**: Git auto-sync (commit/push/pull), records every edit by both humans and Agents. One-click rollback, cross-device sync.
+- **Agent Inspector**: renders Agent operation logs as a filterable timeline for reviewing tool calls in detail.
+- **Git Time Machine**: Git auto-sync (commit/push/pull) keeps human and Agent edits reviewable, with rollback and cross-device sync.
 - **Desktop App**: native macOS/Windows/Linux app with system tray, auto-start, and local process management.
 
 <details>
-<summary><strong>Coming Soon</strong></summary>
+<summary><strong>Roadmap / Experimental</strong></summary>
 
-- [ ] Deep RAG integration: retrieval-augmented generation grounded in your knowledge base for more accurate, context-aware AI responses
-- [ ] ACP / A2A Phase 2: deep multi-agent collaboration with task delegation, shared context, and workflow chaining
-- [ ] Experience Compiler: auto-extract corrections and preferences from agent interactions into reusable Skills/SOPs
-- [ ] Knowledge Health Dashboard: visualize cognitive compound metrics — rules accumulated, agent reuse count, knowledge freshness
+- **Retrieval depth**: richer knowledge retrieval and ranking on top of the current local knowledge base and search surfaces.
+- **Agent coordination**: deeper ACP/A2A collaboration once durable mailbox, task-board, permission, and artifact contracts are ready.
+- **Experience capture**: more guided ways to turn reviewed conversations, corrections, and preferences into reusable Skills/SOPs.
+- **Knowledge health**: lightweight signals for stale notes, reused rules, and knowledge that deserves review.
 
 </details>
 
@@ -255,7 +256,7 @@ graph LR
     H["👤 Human<br/><sub>thinks · reviews · evolves</sub>"]
     M[("📚 MindOS")]
     A["🤖 Agent<br/><sub>executes · retrospects · extracts SOPs</sub>"]
-    EXT["🌐 All Agents"]
+    EXT["🌐 Connected Agents"]
 
     H -- "ideas & feedback" --> M
     M -- "context & insights" --> H
@@ -269,13 +270,15 @@ graph LR
     style EXT fill:#64748b,stroke:#475569,color:#fff,stroke-dasharray:5 5
 ```
 
-> **Both sides evolve.** Humans gain new insights from accumulated knowledge; Agents extract SOPs and get smarter. MindOS sits at the center — the shared second brain that grows with every interaction.
+> **Both sides evolve.** Humans gain new insights from accumulated knowledge; Agents reuse reviewed context and SOPs. MindOS sits at the center — the shared second brain that improves as knowledge accumulates.
 
 ---
 
 ## 🤝 Supported Agents
 
-> The current registry supports 26 Agents. Full list with MCP config paths and manual setup: **[docs/en/supported-agents.md](docs/en/supported-agents.md)**
+MindOS supports the built-in MindOS Agent plus a growing set of CLI, IDE, and extension-based agents. Full list with MCP config paths, Skill paths, and troubleshooting notes: **[docs/en/supported-agents.md](docs/en/supported-agents.md)**
+
+Common examples:
 
 | Agent | MCP | Skills |
 |:------|:---:|:------:|
@@ -298,26 +301,22 @@ graph LR
 
 ```bash
 MindOS/
-├── packages/web/         # Next.js 16 frontend — browse, edit, and interact with AI
-├── packages/desktop/     # Electron desktop client
-├── packages/mobile/      # Expo mobile app
+├── packages/mindos/      # Published runtime package, CLI, server, protocols, and core product logic
+├── packages/web/         # Next.js frontend for the local knowledge workbench
+├── packages/desktop/     # Electron desktop client and local process management
 ├── packages/browser-extension/ # Web Clipper browser extension
-├── packages/desktop-tauri/     # Tauri desktop spike
-├── packages/mindos/      # Published product package: runtime facade, protocols, foundation/knowledge internals, CLI kernel
-├── packages/retrieval/   # Optional retrieval stack: search, vector, indexer, API
-├── skills/           # MindOS Skills (`mindos`, `mindos-zh`) — workflow guides for Agents
-├── templates/        # Preset templates (`en/`, `zh/`, `empty/`) — copied to knowledge base on onboard
-├── scripts/          # Setup wizard and helper scripts
+├── packages/retrieval/   # Optional retrieval/search/vector adapters
+├── packages/mobile/      # Expo mobile app
+├── skills/               # Packaged MindOS Skills for agents
+├── templates/            # Starter knowledge-base templates
+├── docs/                 # Public user and contributor docs
 └── README.md
 
 ~/.mindos/            # User data directory (outside project, never committed)
 ├── config.json       # All configuration (AI keys, port, auth token, sync settings)
-├── sync-state.json   # Sync state (last sync time, conflicts)
 └── mind/             # Your private knowledge base (default path, customizable on onboard)
 ```
 
-> In v1, top-level `app/`, `apps/`, `mcp/`, `desktop/`, `mobile/`, `browser-extension/`, `desktop-tauri/`, and root `bin/` are no longer source roots. The repo root is a private monorepo orchestrator. `packages/mindos` is the OpenCode-style published product package (`@geminilight/mindos`): its repo CLI entry is `packages/mindos/bin/cli.js`, installed npm users still receive the package-relative `bin/cli.js`, and the prebuilt local Web runtime is published as `_standalone/`. `packages/web` remains the only Web source tree and is not copied into the npm tarball. During `npm pack` / publish, `packages/mindos/_standalone`, `packages/mindos/packages`, `packages/mindos/scripts`, `packages/mindos/assets`, `packages/mindos/skills`, and `packages/mindos/templates` may be generated as staging output; they are ignored, excluded from pnpm workspace discovery, and cleaned by `pnpm run clean:product-stage` or product `postpack`.
-
 ## ⌨️ CLI Commands
 
 > Full command reference: **[docs/en/cli-commands.md](docs/en/cli-commands.md)**
@@ -325,31 +324,29 @@ MindOS/
 | Command | Description |
 | :--- | :--- |
 | **Core** | |
-| `mindos onboard` | Interactive setup (config, template, start mode) |
-| `mindos start` | Start Web + MCP services |
-| `mindos start --daemon` | Start as background OS service (macOS/Linux only) |
-| `mindos stop` / `restart` | Stop or restart running processes |
-| `mindos dev` | Start in dev mode |
-| `mindos build` | Build for production |
-| `mindos status` | Show service status overview |
+| `mindos` | Open the MindOS AI Agent REPL |
+| `mindos onboard` / `mindos init` | Interactive setup (config, template, start mode) |
+| `mindos onboard --install-daemon` | Onboard and install the background service on macOS/Linux |
+| `mindos start` / `mindos start --daemon` | Start Web + MCP services in foreground or as a background service |
+| `mindos stop` / `mindos restart` | Stop or restart running processes |
+| `mindos status` | Show service status overview (`--json` supported) |
 | `mindos open` | Open Web UI in browser |
 | **Knowledge** | |
-| `mindos file <sub>` | File operations (list, read, create, delete, search) |
-| `mindos space <sub>` | Space management (list, create, info) |
-| `mindos search "<query>"` | Search knowledge base |
-| `mindos` | Open the MindOS AI Agent REPL |
+| `mindos file <sub>` / `mindos space <sub>` | File and space operations for the local knowledge base |
+| `mindos search "<query>"` | Search the knowledge base |
 | `mindos "<task>"` / `mindos -p "<task>"` | Run the MindOS AI Agent with your knowledge base |
-| `mindos agent <sub>` | Explicit agent command and management (list, info) |
-| `mindos channel` / `mindos feishu-ws` | IM channel config and Feishu long connection |
+| `mindos agent <sub>` | Agent command and management helpers (`list`, `info`, scripted runs) |
 | `mindos api <METHOD> <path>` | Raw API passthrough for developers/agents |
 | **MCP & Config** | |
-| `mindos mcp install` | Auto-install MCP config into your Agent |
-| `mindos token` | Show auth token and MCP config |
-| `mindos config <sub>` | View/update config (show, set, validate) |
+| `mindos mcp` | Start the MCP server only |
+| `mindos mcp install` | Install or repair MCP config and the MindOS Skill for your Agent |
+| `mindos doctor agents [name]` | Verify agent-facing MCP, command, and Skill readiness |
+| `mindos token` | Show auth token and per-agent MCP config snippets |
+| `mindos config <sub>` | View/update config (`show`, `set`, `validate`) |
 | `mindos sync` | Show sync status (init, now, conflicts, on/off) |
-| `mindos gateway <sub>` | Manage background service on macOS/Linux (install, start, stop) |
-| `mindos doctor` | Health check |
-| `mindos update` | Update to latest version |
+| `mindos gateway <sub>` | Manage background service on macOS/Linux (`install`, `start`, `stop`, `logs`) |
+| `mindos doctor` / `mindos logs` | Health check and local service logs |
+| `mindos update` / `mindos uninstall` | Update or remove MindOS |
 
 **Main keyboard shortcuts:** `⌘K` Search · `⌘/` AI Assistant · `E` Edit · `⌘S` Save · `Esc` Close.
 
@@ -379,6 +376,7 @@ Join our WeChat group for early access, feedback, and AI workflow discussions:
 <a href="https://github.com/zjpyb"><img src="https://github.com/zjpyb.png" width="60" style="border-radius:50%" alt="zjpyb" /></a>
 <a href="https://github.com/huzhangyi"><img src="https://github.com/huzhangyi.png" width="60" style="border-radius:50%" alt="huzhangyi" /></a>
 <a href="https://github.com/zenozhg"><img src="https://github.com/zenozhg.png" width="60" style="border-radius:50%" alt="Zeno Zhang" /></a>
+<a href="https://github.com/octo-patch"><img src="https://github.com/octo-patch.png" width="60" style="border-radius:50%" alt="Octopus" /></a>
 
 ### 🙏 Acknowledgements
 

package/README_zh.md

@@ -24,7 +24,7 @@
   <a href="https://github.com/GeminiLight/MindOS/releases?q=clipper"><img src="https://img.shields.io/badge/网页剪藏-安装-c8873a.svg?style=for-the-badge" alt="网页剪藏"></a>
 </p>
 
-MindOS 是你思考的地方，也是 AI Agent 行动的起点——一个你和所有 AI 共享的本地知识库。**每次思考，都在生长。**
+MindOS 是一个本地优先的知识库，用来在你和常用 AI Agent 之间共享长期上下文。**在一处思考，让 Agent 带着共同上下文行动。**
 
 ---
 
@@ -78,7 +78,7 @@ MindOS 是你思考的地方，也是 AI Agent 行动的起点——一个你和
 </table>
 
 > [!IMPORTANT]
-> **⭐ 一键安装：** 把这句话发给你的 Agent（Claude Code、Cursor 等），自动完成全部安装：
+> **⭐ Agent 辅助安装：** 把这句话发给你的 Agent（Claude Code、Cursor 等），安装 MindOS、MCP 和 Skills：
 > ```
 > 帮我从 https://github.com/GeminiLight/MindOS 安装 MindOS，包含 MCP 和 Skills，使用中文模板。
 > ```
@@ -96,14 +96,15 @@ MindOS 是你思考的地方，也是 AI Agent 行动的起点——一个你和
 
 ## 📢 最新动态
 
-| 日期 | 更新 |
+MindOS 现在围绕几个稳定的产品模块展开，而不是一组零散安装脚本：
+
+| 模块 | 当前已上线 |
 |------|------|
-| 2026-04 | **协议 runtime 内聚** — ACP 与 MCP 源码已迁入 `packages/mindos/src/protocols/*`；发布包只带自包含的 `dist/protocols/*` bundle，不再嵌套协议 workspace package。 |
-| 2026-04 | **产品主包落地** — MindOS 采用 OpenCode 式核心包：`packages/mindos`（`@geminilight/mindos`）承载产品 runtime facade 与 CLI kernel 边界；Web/CLI 围绕它做适配。 |
-| 2026-04 | **OpenCode 式 workspace 扁平化** — 所有源码 workspace 已统一到 `packages/`：`packages/web`、`packages/desktop`、`packages/mobile`、`packages/mindos`，以及可选 `packages/retrieval/*` adapters。npm tarball / CLI / MCP / Web standalone / Desktop / Mobile 均已验证。 |
-| 2026-04 | **发布包清理** — 旧顶层 `app/`、`mcp/`、`desktop/`、`mobile/` 源码根已移除，发布清单精确排除测试、缓存与 nested `node_modules`。 |
-| 2026-04 | **网页剪藏插件** — 浏览器扩展，一键剪藏任意网页到 MindOS 收集箱。[安装 →](packages/browser-extension/) |
-| 2026-04 | **PDF & 图片导入** — 拖拽 PDF 和图片到 MindOS，自动转换为可搜索的 Markdown。 |
+| **本地知识工作台** | 桌面端 / Web / CLI 共同管理本地 Markdown 知识库，支持浏览、编辑、导入、搜索与整理。 |
+| **MCP + Skills 桥接** | 一条命令安装或修复 MCP 配置与随包发布的 MindOS Skill，再用 `mindos doctor agents` 验证 Agent 实际可用状态。 |
+| **原生 Agent Runtime** | 支持可恢复会话、取消、重连/重新附着，以及在 runtime 暴露能力时记录可审查的输出历史。 |
+| **工作流与审计面** | YAML 工作流、Agent Inspector、run ledger、反向链接、知识图谱与导入进度，让人类审查保留在流程里。 |
+| **跨平台交付** | 可通过桌面端 Release 或 `npm install -g @geminilight/mindos@latest` 安装；npm 包内置 CLI 与预构建本地 Web runtime。 |
 
 ## 🧠 人机共享心智
 
@@ -111,24 +112,24 @@ MindOS 是你思考的地方，也是 AI Agent 行动的起点——一个你和
 
 **1. 全局同步 — 打破记忆割裂**
 
-切换工具带来上下文割裂，个人深度背景散落各处，导致知识无法复用。**MindOS 内置 MCP Server，所有 Agent 零配置直连核心知识库。项目记忆与 SOP 仅需一处记录，全局赋能所有 AI 工具。**
+切换工具带来上下文割裂，个人深度背景散落各处，导致知识无法复用。**MindOS 内置 MCP Server，并随包提供 Skills，让受支持的 Agent 能共享访问核心知识库。项目记忆与 SOP 只需记录一次，就能在多个 AI 工具里复用。**
 
 **2. 透明可控 — 消除记忆黑箱**
 
-Agent 记忆锁在黑箱中，推理无法审查，错误极难纠正且幻觉持续累积。**MindOS 将每次检索与执行沉淀为本地纯文本，提供完整的审查干预界面，人类拥有绝对的心智纠偏权，随时校准 Agent 行为。**
+Agent 记忆锁在黑箱中，推理难以审查，错误也难以纠正。**MindOS 以本地纯文本保存知识库，并提供 Agent 运行、文件变更和关键工具活动的审查入口，让你能看见并调整 Agent 所依赖的上下文。**
 
 **3. 共生演进 — 经验回流为指令**
 
-反复表达偏好但新对话又从零开始，思考未变成 AI 的能力与自身的方法论。**MindOS 将每次思考自动沉淀为知识库。在交互中厘清想法与标准，下次应对更默契拒绝重复犯错，认知随每次沉淀变得锐利。**
+反复表达偏好但新对话又从零开始，思考没有变成可复用的方法论。**MindOS 帮你把经过审查的对话、纠正和项目标准沉淀为笔记、Skills 与 SOP，让后续 Agent 工作从更好的上下文起步。**
 
-> **底层原则：** 默认本地优先，全部数据以本地纯文本保存，兼顾隐私、主权与性能。
+> **底层原则：** 默认本地优先，知识库内容以本地纯文本保存，兼顾隐私、主权与性能。
 
 ---
 
 ## 🚀 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 3 步](#3-通过-mindos-agent-注入你的个人心智)：
+> **用 Agent 辅助安装：** 将以下指令粘贴到任意支持 MCP 的 Agent（Claude Code、Cursor 等），安装 MindOS、MCP 和 Skills，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。
 > ```
@@ -186,18 +187,18 @@ mindos open
 
 ### 4. 让任意 Agent 可用（MCP + Skills）
 
-**MCP**（连接能力）— 一条命令自动安装：
+**MCP + Skills**（连接能力 + 工作流能力）— 一条命令同时安装或修复：
 
 ```bash
 mindos mcp install        # 交互式
 mindos mcp install -g -y  # 一键全局安装
 ```
 
-**Skills**（工作流能力）— 根据语言偏好选择其一：
+验证 Agent 实际可用状态：
 
 ```bash
-npx skills add https://github.com/GeminiLight/MindOS --skill mindos -g -y      # 英文
-npx skills add https://github.com/GeminiLight/MindOS --skill mindos-zh -g -y   # 中文
+mindos doctor agents
+mindos doctor agents codex --json
 ```
 
 > 远程配置、手动 JSON 片段、常见误区等详见 **[docs/zh/supported-agents.md](docs/zh/supported-agents.md)**。
@@ -223,28 +224,28 @@ MindOS 的知识不必从零开始——把你已有的信息导入进来：
 
 **Agent 侧**
 
-- **MCP Server + Skills**：stdio + HTTP 双传输，全阵容 Agent 兼容（Claude Code, Cursor, Gemini CLI 等），零配置接入。
-- **ACP / A2A 协议**：Agent 间通信协议，支持 Agent 发现、任务委派与编排。Phase 1 已上线：Agent Card 发现 + JSON-RPC 消息通信。
-- **Workflow 编排**：基于 YAML 的可视化工作流编辑器 + 步骤执行引擎，定义、编辑、运行多步 Agent 工作流。
+- **MCP Server + Skills**：为受支持的 Agent（Claude Code、Cursor、Gemini CLI、Codex 等）提供 stdio + HTTP 传输，并通过 installer/doctor 命令完成安装与修复。
+- **ACP / A2A 基础能力**：提供 Agent 发现、注册表、会话检测与 JSON-RPC 消息等早期协议面；持久 mailbox/task-board 等深层协作仍处于实验阶段。
+- **Workflow 编排**：基于 YAML 的可视化工作流编辑器 + 步骤执行引擎，用于定义、编辑和运行可复用的 Agent 工作流。
 - **结构化模板**：预置 Profile、Workflows、Configurations 等目录骨架，快速冷启动个人 Context。
-- **笔记即指令**：日常笔记天然就是 Agent 可直接执行的高质量指令——无需额外格式转换，写下即可调度。
+- **笔记即指令**：日常笔记可以沉淀为 Agent 可复用、可执行的指令，写下上下文后再调度或持续修订。
 
 **基础设施**
 
 - **安全防线**：Bearer Token 认证、路径沙箱、INSTRUCTION.md 写保护、原子写入。
 - **知识图谱**：动态解析并可视化文件间的引用与依赖关系。
 - **反向链接视图**：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**：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。
-- **Git 时光机**：Git 自动同步（commit/push/pull），记录人类与 Agent 的每次编辑历史，一键回滚，跨设备同步。
+- **Agent 审计面板**：将 Agent 操作日志渲染为可筛选的时间线，用于审查工具调用详情。
+- **Git 时光机**：Git 自动同步（commit/push/pull），让人类与 Agent 的编辑保留可审查历史，支持回滚与跨设备同步。
 - **桌面客户端**：原生 macOS/Windows/Linux 应用，系统托盘、开机自启、本地进程管理。
 
 <details>
-<summary><strong>即将到来</strong></summary>
+<summary><strong>路线图 / 实验能力</strong></summary>
 
-- [ ] 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
-- [ ] ACP / A2A Phase 2：深度多 Agent 协作，支持任务委派、共享上下文、工作流链式执行
-- [ ] 经验编译器：从 Agent 交互中自动提取纠正和偏好，沉淀为可复用的 Skills/SOP
-- [ ] 知识库健康度仪表盘：可视化认知复利指标——已积累规则数、Agent 复用次数、知识新鲜度
+- **检索深度**：在现有本地知识库和搜索能力之上，继续增强检索与排序质量。
+- **Agent 协作**：等 mailbox、task-board、权限和 artifact contract 更稳定后，再推进更深层 ACP/A2A 协作。
+- **经验沉淀**：提供更顺手的方式，把审查后的对话、纠正和偏好转为可复用的 Skills/SOP。
+- **知识健康度**：为过期笔记、被复用的规则和需要复查的知识提供轻量信号。
 
 </details>
 
@@ -255,7 +256,7 @@ graph LR
     H["👤 人类<br/><sub>思考 · 审查 · 进化</sub>"]
     M[("📚 MindOS")]
     A["🤖 Agent<br/><sub>执行 · 复盘 · 提炼 SOP</sub>"]
-    EXT["🌐 所有 Agent"]
+    EXT["🌐 已连接 Agent"]
 
     H -- "想法 & 反馈" --> M
     M -- "上下文 & 洞察" --> H
@@ -269,13 +270,15 @@ graph LR
     style EXT fill:#64748b,stroke:#475569,color:#fff,stroke-dasharray:5 5
 ```
 
-> **双向进化。** 人类从积累的知识中获得新洞察；Agent 提炼 SOP 变得更强。MindOS 居中——随每次交互持续成长的共享第二大脑。
+> **双向进化。** 人类从积累的知识中获得新洞察；Agent 复用经过审查的上下文与 SOP。MindOS 居中——随着知识积累不断改进的共享第二大脑。
 
 ---
 
 ## 🤝 支持的 Agent
 
-> 当前注册表支持 26 个 Agent。完整列表及 MCP 配置路径、手动配置方法：**[docs/zh/supported-agents.md](docs/zh/supported-agents.md)**
+MindOS 支持内置 MindOS Agent，也支持一批 CLI、IDE 和扩展类 Agent。完整列表、MCP 配置路径、Skill 路径与常见问题见：**[docs/zh/supported-agents.md](docs/zh/supported-agents.md)**
+
+常见示例：
 
 | Agent | MCP | Skills |
 |:------|:---:|:------:|
@@ -298,26 +301,22 @@ graph LR
 
 ```bash
 MindOS/
-├── packages/web/         # Next.js 16 前端 — 浏览、编辑、与 AI 交互
-├── packages/desktop/     # Electron 桌面客户端
-├── packages/mobile/      # Expo 移动端
+├── packages/mindos/      # 发布 runtime 包、CLI、server、protocols 与核心产品逻辑
+├── packages/web/         # 本地知识工作台的 Next.js 前端
+├── packages/desktop/     # Electron 桌面端与本地进程管理
 ├── packages/browser-extension/ # Web Clipper 浏览器扩展
-├── packages/desktop-tauri/     # Tauri 桌面 spike
-├── packages/mindos/      # 已发布产品包：runtime facade、protocols、foundation/knowledge 内部模块、CLI kernel
-├── packages/retrieval/   # 可选检索栈：search/vector/indexer/API
-├── skills/           # MindOS Skills（`mindos`、`mindos-zh`）— Agent 工作流指南
-├── templates/        # 预设模板（`en/`、`zh/`、`empty/`）— onboard 时复制到知识库目录
-├── scripts/          # 配置向导与辅助脚本
+├── packages/retrieval/   # 可选检索 / 搜索 / 向量适配器
+├── packages/mobile/      # Expo 移动端
+├── skills/               # 随包发布的 MindOS Skills
+├── templates/            # 知识库初始化模板
+├── docs/                 # 公开用户文档与贡献者文档
 └── README.md
 
 ~/.mindos/            # 用户数据目录（项目外，不会被提交）
 ├── config.json       # 所有配置（AI 密钥、端口、Auth token、同步设置）
-├── sync-state.json   # 同步状态（最后同步时间、冲突文件）
 └── mind/             # 你的私有知识库（默认路径，onboard 时可自定义）
 ```
 
-> v1 之后，顶层 `app/`、`apps/`、`mcp/`、`desktop/`、`mobile/`、`browser-extension/`、`desktop-tauri/` 和根目录 `bin/` 不再是源码入口。仓库根目录是 private monorepo 编排层。`packages/mindos` 是 OpenCode 式发布产品包（`@geminilight/mindos`）：仓库内 CLI 入口是 `packages/mindos/bin/cli.js`，npm 安装后用户仍获得包内相对路径 `bin/cli.js`，预构建本地 Web runtime 以 `_standalone/` 发布。`packages/web` 是唯一 Web 源码，不会复制进 npm tarball。`npm pack` / publish 期间可能生成 `packages/mindos/_standalone`、`packages/mindos/packages`、`packages/mindos/scripts`、`packages/mindos/assets`、`packages/mindos/skills`、`packages/mindos/templates` 等 staging 输出；这些不是源码，已被 git 和 pnpm workspace 排除，可用 `pnpm run clean:product-stage` 或 product `postpack` 清理。
-
 ## ⌨️ CLI 命令
 
 > 完整命令参考：**[docs/zh/cli-commands.md](docs/zh/cli-commands.md)**
@@ -325,31 +324,29 @@ MindOS/
 | 命令 | 说明 |
 | :--- | :--- |
 | **核心** | |
-| `mindos onboard` | 交互式初始化（生成配置、选择模板） |
-| `mindos start` | 前台启动 app + MCP 服务 |
-| `mindos start --daemon` | 以后台 OS 服务方式启动（仅 macOS/Linux） |
-| `mindos stop` / `restart` | 停止或重启服务 |
-| `mindos dev` | 开发模式启动 |
-| `mindos build` | 构建生产版本 |
-| `mindos status` | 查看服务状态概览 |
+| `mindos` | 进入 MindOS AI Agent 交互模式 |
+| `mindos onboard` / `mindos init` | 交互式初始化（配置、模板、启动模式） |
+| `mindos onboard --install-daemon` | 初始化并安装 macOS/Linux 后台服务 |
+| `mindos start` / `mindos start --daemon` | 前台或后台启动 Web + MCP 服务 |
+| `mindos stop` / `mindos restart` | 停止或重启服务 |
+| `mindos status` | 查看服务状态概览（支持 `--json`） |
 | `mindos open` | 在浏览器中打开 Web UI |
 | **知识库** | |
-| `mindos file <sub>` | 文件操作（list, read, create, delete, search） |
-| `mindos space <sub>` | 空间管理（list, create, info） |
+| `mindos file <sub>` / `mindos space <sub>` | 本地知识库文件与空间操作 |
 | `mindos search "<query>"` | 搜索知识库 |
-| `mindos` | 进入 MindOS AI Agent 交互模式 |
 | `mindos "<task>"` / `mindos -p "<task>"` | 基于知识库运行 MindOS AI Agent |
-| `mindos agent <sub>` | 显式 agent 命令与管理（list, info） |
-| `mindos channel` / `mindos feishu-ws` | IM 渠道配置与飞书长连接 |
+| `mindos agent <sub>` | Agent 命令与管理辅助（`list`、`info`、脚本化运行） |
 | `mindos api <METHOD> <path>` | API 透传（开发者/Agent 用） |
 | **MCP 与配置** | |
-| `mindos mcp install` | 自动将 MCP 配置写入 Agent |
-| `mindos token` | 查看认证令牌和 MCP 配置 |
-| `mindos config <sub>` | 查看/修改配置（show, set, validate） |
+| `mindos mcp` | 仅启动 MCP server |
+| `mindos mcp install` | 为 Agent 安装或修复 MCP 配置与 MindOS Skill |
+| `mindos doctor agents [name]` | 验证 Agent 侧 MCP、命令与 Skill 可用状态 |
+| `mindos token` | 查看认证令牌与各 Agent MCP 配置片段 |
+| `mindos config <sub>` | 查看/修改配置（`show`、`set`、`validate`） |
 | `mindos sync` | 同步状态（init, now, conflicts, on/off） |
-| `mindos gateway <sub>` | 管理 macOS/Linux 后台服务（install, start, stop） |
-| `mindos doctor` | 健康检查 |
-| `mindos update` | 更新到最新版本 |
+| `mindos gateway <sub>` | 管理 macOS/Linux 后台服务（`install`、`start`、`stop`、`logs`） |
+| `mindos doctor` / `mindos logs` | 健康检查与本地服务日志 |
+| `mindos update` / `mindos uninstall` | 更新或卸载 MindOS |
 
 **主要快捷键：** `⌘K` 搜索 · `⌘/` AI 助手 · `E` 编辑 · `⌘S` 保存 · `Esc` 关闭。
 
@@ -379,6 +376,7 @@ MindOS/
 <a href="https://github.com/zjpyb"><img src="https://github.com/zjpyb.png" width="60" style="border-radius:50%" alt="zjpyb" /></a>
 <a href="https://github.com/huzhangyi"><img src="https://github.com/huzhangyi.png" width="60" style="border-radius:50%" alt="huzhangyi" /></a>
 <a href="https://github.com/zenozhg"><img src="https://github.com/zenozhg.png" width="60" style="border-radius:50%" alt="Zeno Zhang" /></a>
+<a href="https://github.com/octo-patch"><img src="https://github.com/octo-patch.png" width="60" style="border-radius:50%" alt="Octopus" /></a>
 
 ### 🙏 致谢
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@geminilight/mindos-linux-arm64-musl",
-  "version": "1.1.58",
+  "version": "1.1.60",
   "description": "MindOS runtime package for linux-arm64-musl",
   "type": "module",
   "license": "MIT",

package/runtime-manifest.json

@@ -2,7 +2,7 @@
   "schemaVersion": 1,
   "product": {
     "name": "@geminilight/mindos",
-    "version": "1.1.58"
+    "version": "1.1.60"
   },
   "package": {
     "name": "@geminilight/mindos-linux-arm64-musl",