# Mysti 2.0 BETA: Multi-Agent Coding Workspace for VS Code > Run multiple coding agents in VS Code with @mentions, ordered handoffs, second opinions and Brainstorm synthesis. Mysti 2.0 beta. ## Install Copy the content below into your project: # Mysti 2.0 BETA: Multi-Agent Coding Workspace for VS Code Run multiple coding agents in VS Code with @mentions, ordered handoffs, second opinions and Brainstorm synthesis. Mysti 2.0 beta. > Source reviewed; runtime not tested. This guide is based on the captured repository README; commands have not been executed here. ## Quick Use 1. **Install Mysti** from the [VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=DeepMyst.mysti), or install the reviewed beta VSIX with **Extensions → Install from VSIX**. Choose the pre-release channel when 2.0 BETA is published. 2. **Open a trusted local workspace.** Mysti supports VS Code 1.86 or newer on Windows, macOS, and Linux. Native voice support and provider CLIs have their own host requirements. 3. **Open Mysti and choose your agent.** The setup screen checks installation and authentication and provides the supported install, upgrade, or sign-in action. 4. **Start with a small task**, for example: ```text @claude Explain the main entry point ``` Then add a second agent for an independent opinion: ```text @claude @codex What are your opinions on adding a cache to this API? ``` Or express a dependency between steps: ```text @claude Design a retry policy, then @codex review it for edge cases ``` You only need one configured agent to begin. See [Getting started](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/GETTING-STARTED.md) and [Provider setup](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROVIDERS.md) in the repository. ## Introduction Mysti is an open-source coding workspace (Apache-2.0) that brings several coding agents together inside VS Code. Instead of juggling separate terminals, you address agents in one conversation with explicit tags such as `@claude` and `@codex`, see separately attributed answers, chain work with `then`, and keep drafts, context and permissions in one place. It also ships a Mysti coordinator backed by a DeepMyst account, plus Brainstorm discussion strategies and Proactive context gathering. The README describes the 2.0.0 pre-release checkout, i.e. a beta. 2.0 introduces parallel opinions, cross-agent handoffs, connected tools and Proactive context on top of the single-agent experience. ## What you can do | Need | How Mysti addresses it | | --- | --- | | A direct answer or implementation | Pick a provider, or tag `@claude` / `@codex` | | Independent perspectives | `@claude @codex` with a question or review request; read-only assignments run concurrently and stay separately attributed | | An ordered handoff | `@claude Write the parser, then @codex review it` — the reviewer receives the completed result; a failed dependency blocks the dependent step | | A defined perspective | Catalog roles such as `@claude:critic` or `@codex:reviewer` | | A structured discussion | Brainstorm strategies: Quick, Debate, Red Team, Perspectives, Delphi | | A coordinating agent | The Mysti coordinator, with optional configured local execution and delegation behind permission gates | | Visual iteration | Canvas and Visual Test; browser setup may be required | | Reusable project practices | Personas, skills and roles | Concurrency follows the work: independent advisory requests may overlap, while writers sharing a workspace run serially. Model, effort and Ultracode controls are shown where the selected provider supports them. ## Contributor build commands The README lists these commands for working on Mysti itself: ```sh npm ci npm run typecheck npm test npm run watch ``` Use the Node version in `.nvmrc`; press **F5** to launch the extension development host. See `CONTRIBUTING.md` for provider contracts, browser tests and contribution expectations. ## Prerequisites - VS Code 1.86 or newer on Windows, macOS or Linux. - A trusted local workspace to open. - At least one configured agent. Provider families listed by the README: | Provider family | Available adapters | | --- | --- | | Coding agents | Claude Code, OpenAI Codex, Google Gemini CLI, GitHub Copilot CLI, Cline, Cursor | | Additional coding tools | OpenClaw, OpenCode, Qwen Code, Hermes, Continue, Kimi Code | | Local / API endpoints | Ollama, LocalAI, OpenRouter, MiniMax | | Mysti coordinator | DeepMyst account, with optional configured local execution and delegation | - For the coordinator, connected tools and cloud responsibilities: a DeepMyst account. - For the contributor workflow: Node (version pinned in `.nvmrc`), npm, and optionally a browser for Visual Test. ## Permissions, data flow and limitations - **Credentials are separate.** Mysti itself uses your DeepMyst account; CLI agents use their own credentials. Provider subscriptions, API access and usage charges vary by provider. - **Prompt transmission.** Provider requests transmit the prompt and supplied context to the configured provider. DeepMyst features use your connected account. - **Tool connections are opt-in and scoped.** Connected apps and MCP servers go through DeepMyst's Composio and Smithery catalogs; browse and authorize in DeepMyst, then manage connections from Mysti. Catalog availability and tool permissions depend on the service and your account. Local CLI access is opt-in for supported adapters. A connected app is not automatically a Proactive monitoring source. - **Proactive limits.** Cloud monitoring requires DeepMyst sign-in, a supported connection and an explicitly configured responsibility. Coverage is bounded, and evidence does not prove who owns a task. Local Git watches need the editor running. - **Checkpoints are local.** A checkpoint cannot undo an external action. - **Capability variance.** Model availability, effort, Ultracode, attachments, native approvals and authentication are provider-specific. An adapter being included is not a claim that every provider/version/account combination has been live-tested. - **Beta status.** 2.0 is a pre-release package; the Marketplace badge shows the currently published version. Review the release notes for migration and known limitations before adopting it for important work. - **Telemetry** follows VS Code's telemetry setting; see `.github/SECURITY.md` for the security policy. - The demo recordings in the README use shipped webviews with deterministic sample responses; they are not live provider runs or performance benchmarks. ## FAQ **Do I need all 16 provider adapters configured?** No. The README states you only need one configured agent to begin; more agents enable parallel opinions, handoffs and Brainstorm. **What is the difference between a second opinion and Brainstorm?** A second opinion resends the original question to another configured agent so you can compare separately attributed answers. Brainstorm pairs two configured agents with a discussion strategy (Quick, Debate, Red Team, Perspectives, Delphi) and keeps individual reasoning visible alongside the synthesis. **What happens if one agent in a workflow fails?** A failed participant stays visible; Mysti does not invent its answer or ask another agent to impersonate it, and a failed dependency blocks its dependent step. **Do agents run in parallel?** Independent read-only assignments run concurrently within the configured limit; work that may edit a shared workspace runs serially. **Are the GIFs and video in the README live agent runs?** No. They use shipped webviews with deterministic sample responses to demonstrate interaction and presentation, not live provider runs or benchmarks. ## Source and thanks - Repository: [DeepMyst/Mysti](https://github.com/DeepMyst/Mysti), README at commit `0df2961a227d53308093894d586944e9f656a0a8`. - License: [Apache License 2.0](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/LICENSE) (© 2025 DeepMyst Inc). - Built by [DeepMyst](https://www.deepmyst.com/mysti) and the Mysti community. Provider names and trademarks belong to their respective owners. - Additional docs referenced by the README: `docs/GETTING-STARTED.md`, `docs/MENTIONS.md`, `docs/PROVIDERS.md`, `docs/FEATURES.md`, `docs/BRAINSTORM.md`, `docs/PERSONAS-AND-SKILLS.md`, `docs/PROACTIVE.md`, `docs/ARCHITECTURE.md`, `docs/releases/2.0-beta/RELEASE-NOTES.md`, `CONTRIBUTING.md`. - More ways to get involved: [Report a bug](https://github.com/DeepMyst/Mysti/issues/new/choose) · [Discuss an idea](https://github.com/DeepMyst/Mysti/issues) · [View the roadmap](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/plans/README.md). ## Complete upstream material
Install Mysti · Get started · 2.0 BETA release notes · Contribute
 # Mysti 2.0 BETA **An open-source coding workspace that brings your agents together inside VS Code.** Ask Claude Code and Codex for independent opinions. Give a specialist a specific task. Hand completed work to another agent for review. Keep their responses, context, permissions, and progress in one conversation. Mysti connects the coding tools you choose—CLI agents, local model servers, and the DeepMyst-powered Mysti coordinator. Use one agent when that's enough; bring in more when another perspective helps. [](https://github.com/DeepMyst/Mysti/actions/workflows/ci.yml) [](https://marketplace.visualstudio.com/items?itemName=DeepMyst.mysti) [](LICENSE) > **2.0 is a beta.** This checkout prepares the 2.0.0 pre-release package. The Marketplace badge shows the currently published version. See [release status and limitations](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/releases/2.0-beta/RELEASE-NOTES.md) before adopting it for important work. **16 provider adapters + the Mysti coordinator · Parallel opinions · Cross-agent handoffs · Connected tools · Proactive context** ## See it work [**Watch the product tour →**](https://github.com/DeepMyst/Mysti/blob/main/docs/releases/2.0-beta/mysti-2-beta-tour.mp4) The recordings below use the **shipped webviews with deterministic sample responses**. They demonstrate interaction and presentation; they are not live provider runs or performance benchmarks. [Capture details and video downloads](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/releases/2.0-beta/README.md). ### Get a second opinion without rewriting the question An answer is a starting point. On answers with a usage or session receipt, click **Second opinion**, choose another configured agent, and Mysti sends it the original question. Compare separately attributed answers in the same conversation; use Brainstorm when you want a discussion and synthesis.  ### Ask two agents. Hear from both. ```text @claude @codex What are your opinions on adding a cache to this API? ``` Explicit tags determine who does the work. Independent read-only assignments run concurrently within the configured limit. Each agent has its own live card and attributed answer—even if one of them is already your selected provider.  A failed participant stays visible. Mysti does not invent its answer or quietly ask the other agent to impersonate it. ### Turn tags into a workflow ```text @claude Design a retry policy, then @codex review it for edge cases ``` Use **then** to make the next assignment depend on the previous result. Claude designs the policy; Codex receives that result to review. Progress cards show the current step and who runs next. A failed dependency blocks its dependent step.  Add roles such as `@claude:critic` or `@codex:reviewer` to make the perspective explicit. Independent advisory work can run in parallel; work that may edit a shared workspace runs serially. [Assignment and workflow guide](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/MENTIONS.md). ### Switch agents without losing your draft Choose from **16 registered provider adapters**, plus the **Mysti coordinator**. Move between coding CLIs, local model servers, and API providers from the composer. Your draft stays in place, and model controls follow the selected provider's capabilities.  [See all providers and setup requirements](#bring-your-preferred-agents). ### Keep model choices close to your work Choose the model, adjust effort, and toggle Ultracode where the provider supports it. The model menu and action menu share the same settings. Dictation adds text to your draft for review before you send.  ### Bring your tools into the conversation Connect apps and MCP servers through **DeepMyst's Composio and Smithery catalogs**. Bring code, team discussions, email, and calendar context within reach of your agents through the tools you authorize. Browse and authorize in DeepMyst, then manage connections from Mysti.  The recording shows sample authorized connections, not a live OAuth flow. Catalog availability and tool permissions depend on the service and your account; local CLI access is opt-in for supported adapters. A connected app is not automatically a Proactive monitoring source. ### Get context before you start Proactive brings selected local Git changes and DeepMyst-connected GitHub/Slack evidence into an inbox. The **Before you start** view helps you inspect relevant evidence for a responsibility, including source links and freshness information.  Cloud monitoring requires DeepMyst sign-in, a supported connection, and an explicitly configured responsibility. Coverage is bounded; evidence does not prove who owns a task. [Proactive setup and limits](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROACTIVE.md). ### Work through a decision with Brainstorm Choose two configured agents and a discussion strategy. **Quick** collects independent analyses and synthesizes a response; Debate, Red Team, Perspectives and Delphi provide other ways to examine a proposal. Keep the individual reasoning visible alongside the synthesis.  [Choose a Brainstorm strategy](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/BRAINSTORM.md). ## More ways to move the work forward | Feature | Why it matters | Explore | | --- | --- | --- | | **Brainstorm** | Let two agents analyze, debate or challenge a proposal, then synthesize a response. | [Discussion strategies](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/BRAINSTORM.md) | | **Canvas + Visual Test** | Iterate on visual artifacts and inspect a running app alongside the code. | [Visual workflows](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md) | | **Personas, skills and roles** | Reuse project guidance and specialist perspectives instead of repeating instructions. | [Team practices](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PERSONAS-AND-SKILLS.md) | | **Checkpoints + visible approvals** | Inspect requested actions and recover supported local changes as you iterate. Checkpoints do not undo external actions. | [Access and lifecycle](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md#permission-system) | | **Dictation + context tools** | Speak a draft, mention files, or attach diagnostics and Git context before sending. | [Composer and context](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md#chat-interface) | ## One workspace, several ways to work | When you need… | Use… | What happens | | --- | --- | --- | | A direct answer or implementation | A selected provider, or `@claude` / `@codex` | The named agent handles the request with your access settings. | | Independent perspectives | `@claude @codex` with a question or review request | Read-only assignments run concurrently; results remain separately attributed. | | An ordered handoff | `@claude Write the parser, then @codex review it` | The reviewer receives the completed result as reference material. A failed dependency blocks the handoff. | | A defined perspective | `@claude:critic @codex:reviewer` | Catalog roles shape each response and restrict its access. | | A structured discussion | Brainstorm | Choose a team and discussion strategy, then follow its rounds and synthesis. | | A coordinating agent | Mysti | A DeepMyst-backed coordinator can answer, use configured tools, and delegate through permission gates. | | Visual iteration | Canvas and Visual Test | Work with visual artifacts and inspect a running app; browser setup may be required. | | Project-specific practices | Personas, skills, and roles | Reuse your team's guidance with explicit trust and access boundaries. |  Parallelism follows the work: independent advisory requests may overlap; writers share a workspace and run serially. Use **then** for an explicit dependency. [Assignment syntax and troubleshooting](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/MENTIONS.md). ## Get started 1. **Install Mysti** from the [VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=DeepMyst.mysti), or install the reviewed beta VSIX with **Extensions → Install from VSIX**. Choose the pre-release channel when 2.0 BETA is published. 2. **Open a trusted local workspace.** Mysti supports VS Code 1.86 or newer on Windows, macOS, and Linux. Native voice support and provider CLIs have their own host requirements. 3. **Open Mysti and choose your agent.** The setup screen checks installation and authentication and provides the supported install, upgrade, or sign-in action. Mysti itself uses your DeepMyst account; CLI agents use their own credentials. 4. **Start with a small task.** Try `@claude Explain the main entry point`, then add `@codex` for a second opinion when both are configured. You only need **one** configured agent to begin. Provider subscriptions, API access, usage charges, and supported models vary. [Getting started](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/GETTING-STARTED.md) · [Provider setup](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROVIDERS.md). ## Bring your preferred agents | Provider family | Available adapters | | --- | --- | | Coding agents | Claude Code, OpenAI Codex, Google Gemini CLI, GitHub Copilot CLI, Cline, Cursor | | Additional coding tools | OpenClaw, OpenCode, Qwen Code, Hermes, Continue, Kimi Code | | Local / API endpoints | Ollama, LocalAI, OpenRouter, MiniMax | | Mysti coordinator | DeepMyst account, with optional configured local execution and delegation | The UI follows each adapter's capabilities. Model availability, effort, Ultracode, attachments, native approvals, and authentication are provider-specific. An adapter being included is not a claim that every provider/version/account combination has been live-tested. [Compatibility and setup](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROVIDERS.md). ## Your context and your controls - **Context where it belongs:** mention workspace files, inspect diagnostics with `@problems`, or add a Git summary with `@git`. - **Visible authority:** read-only and approval modes, agent-specific tool activity, and cancellation across active assignments. - **Continuity:** conversation history, context compaction, and workspace checkpoints for supported local changes. - **Explicit connections:** choose the DeepMyst tools and sources used by connected features. Cloud responsibilities continue on the backend; local Git watches need the editor running. Provider requests transmit the prompt and supplied context to the configured provider. DeepMyst features use your connected account. Telemetry follows VS Code's telemetry setting; see the [security policy](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/.github/SECURITY.md) and feature-specific guides. A local checkpoint cannot undo an external action. ## Explore the documentation | Guide | What it covers | | --- | --- | | [Start here](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/GETTING-STARTED.md) | Install, configure, send your first task, troubleshoot | | [Agent assignments](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/MENTIONS.md) | Tags, roles, parallel reviews, dependent handoffs | | [Providers](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROVIDERS.md) | Installation, authentication, models, capability differences | | [Feature guide](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md) | Chat, menus, dictation, Canvas, permissions and more | | [Brainstorm](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/BRAINSTORM.md) | Team discussion and synthesis | | [Personas and skills](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PERSONAS-AND-SKILLS.md) | Reusable project practices | | [Proactive](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROACTIVE.md) | Responsibilities, evidence, notification and coverage limits | | [Architecture](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/ARCHITECTURE.md) | Host, webview, routing and provider boundaries | | [Release notes](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/releases/2.0-beta/RELEASE-NOTES.md) | What's new, migration, known limitations | | [Contributing](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/CONTRIBUTING.md) | Development, meaningful tests, demos and release work | **Translations:** [العربية](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ar.md) · [中文](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.zh-CN.md) · [日本語](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ja.md) · [한국어](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ko.md) · [Español](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.es.md) · [Português](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.pt-BR.md) · [Deutsch](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.de.md) · [Français](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.fr.md) · [Türkçe](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.tr.md) · [Русский](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ru.md). Community translations may describe an earlier release; this README is the current 2.0 BETA reference. ## Build with us Mysti is Apache-2.0 licensed. Useful contributions include reproducible routing bugs, native-provider compatibility checks, accessible UI improvements, and tested documentation corrections. ```sh npm ci npm run typecheck npm test npm run watch ``` Use the Node version in `.nvmrc`; press **F5** to launch the extension development host. See [CONTRIBUTING.md](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/CONTRIBUTING.md) for provider contracts, browser tests, and contribution expectations. [Report a bug](https://github.com/DeepMyst/Mysti/issues/new/choose) · [Discuss an idea](https://github.com/DeepMyst/Mysti/issues) · [View the roadmap](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/plans/README.md) ## Contributors Thanks to everyone who has helped make Mysti better!安装 Mysti · 快速上手 · 2.0 BETA 发布说明 · 参与贡献
 # Mysti 2.0 BETA **一个开源编码工作区,在 VS Code 内把你的智能体汇聚到一起。** 向 Claude Code 和 Codex 征求独立意见。给专家智能体分配具体任务。把完成的工作交给另一个智能体评审。将它们的回复、上下文、权限和进度保留在同一个会话中。 Mysti 连接你选择的编码工具——CLI 智能体、本地模型服务器,以及由 DeepMyst 驱动的 Mysti 协调器。一个智能体够用时就用一个;需要另一种视角时再引入更多。 [](https://github.com/DeepMyst/Mysti/actions/workflows/ci.yml) [](https://marketplace.visualstudio.com/items?itemName=DeepMyst.mysti) [](LICENSE) > **2.0 是 beta 版本。** 本次检出内容用于准备 2.0.0 预发布包。Marketplace 徽章显示的是当前已发布的版本。在将其用于重要工作前,请先阅读[发布状态与限制](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/releases/2.0-beta/RELEASE-NOTES.md)。 **16 个 provider 适配器 + Mysti 协调器 · 并行意见 · 跨智能体交接 · 已连接工具 · 主动上下文** ## 看看它如何工作 [**观看产品导览 →**](https://github.com/DeepMyst/Mysti/blob/main/docs/releases/2.0-beta/mysti-2-beta-tour.mp4) 下面的录屏使用的是**随附的 webview 与确定性的示例响应**。它们展示的是交互与呈现方式;并非真实 provider 运行,也不是性能基准。[录屏细节与视频下载](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/releases/2.0-beta/README.md)。 ### 不必重写问题就能获得第二意见 一个回答只是起点。对于带有用量或会话回执的回答,点击 **Second opinion**,选择另一个已配置的智能体,Mysti 会把原始问题发送给它。在同一个会话中比较各自独立署名的回答;当你需要讨论与综合时,使用 Brainstorm。  ### 向两个智能体提问。听取双方回答。 ```text @claude @codex What are your opinions on adding a cache to this API? ``` 显式标签决定由谁完成工作。独立的只读任务在配置的限制内并发执行。每个智能体都有自己的实时卡片和署名回答——即使其中之一已经是你当前选中的 provider。  失败的参与者仍保持可见。Mysti 不会编造它的回答,也不会悄悄让另一个智能体冒充它。 ### 把标签变成工作流 ```text @claude Design a retry policy, then @codex review it for edge cases ``` 使用 **then** 让下一个指派依赖上一步的结果。Claude 设计策略;Codex 接收该结果进行评审。进度卡片会显示当前步骤以及下一个由谁执行。依赖失败会阻塞其后续步骤。  添加诸如 `@claude:critic` 或 `@codex:reviewer` 的角色,让视角更加明确。独立的咨询类工作可以并行运行;可能编辑共享工作区的工作串行运行。[指派与工作流指南](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/MENTIONS.md)。 ### 切换智能体而不丢失草稿 从 **16 个已注册的 provider 适配器**,再加上 **Mysti 协调器**中进行选择。在编写器中即可在编码 CLI、本地模型服务器与 API provider 之间切换。你的草稿保持不变,模型控件会跟随所选 provider 的能力。  [查看所有 provider 与安装要求](#bring-your-preferred-agents)。 ### 让模型选择贴近你的工作 选择模型、调整 effort,并在 provider 支持时切换 Ultracode。模型菜单与操作菜单共用同一套设置。听写会在你发送前把文本加入草稿供审阅。  ### 把你的工具带入对话 通过 **DeepMyst 的 Composio 与 Smithery 目录**连接应用与 MCP 服务器。借助你授权的工具,让代码、团队讨论、邮件与日历上下文触手可及。在 DeepMyst 中浏览并授权,然后在 Mysti 中管理连接。  录屏展示的是示例授权连接,并非真实 OAuth 流程。目录可用性与工具权限取决于该服务与你的账号;对受支持的适配器而言,本地 CLI 访问是可选开启的。已连接的应用不会自动成为 Proactive 监控来源。 ### 开始之前先获取上下文 Proactive 会把选定的本地 Git 变更以及 DeepMyst 已连接的 GitHub/Slack 证据汇入收件箱。**Before you start** 视图帮助你检查与某项职责相关的证据,包括来源链接与新鲜度信息。  云端监控需要登录 DeepMyst、受支持的连接,以及显式配置的职责。覆盖范围有限;证据并不能证明谁拥有某项任务。[Proactive 设置与限制](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROACTIVE.md)。 ### 用 Brainstorm 推进一项决策 选择两个已配置的智能体以及一种讨论策略。**Quick** 收集独立分析并综合出一个回答;Debate、Red Team、Perspectives 与 Delphi 提供其他审视提案的方式。在综合结果旁保留各自的推理过程。  [选择一种 Brainstorm 策略](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/BRAINSTORM.md)。 ## 更多推进工作的方式 | 功能 | 为何重要 | 了解详情 | | --- | --- | --- | | **Brainstorm** | 让两个智能体分析、辩论或质疑一个提案,然后综合出回答。 | [讨论策略](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/BRAINSTORM.md) | | **Canvas + Visual Test** | 迭代可视化产物,并在代码旁检查正在运行的应用。 | [可视化工作流](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md) | | **Personas、skills 与 roles** | 复用项目指导与专家视角,而不必重复指令。 | [团队实践](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PERSONAS-AND-SKILLS.md) | | **检查点 + 可见审批** | 在迭代过程中检查请求的操作,并恢复受支持的本地变更。检查点无法撤销外部操作。 | [访问与生命周期](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md#permission-system) | | **听写 + 上下文工具** | 在发送前说出草稿、提及文件,或附加诊断信息与 Git 上下文。 | [编写器与上下文](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md#chat-interface) | ## 一个工作区,多种工作方式 | 当你需要…… | 使用…… | 会发生什么 | | --- | --- | --- | | 直接的回答或实现 | 选定的 provider,或 `@claude` / `@codex` | 被指定的智能体按你的访问设置处理请求。 | | 独立视角 | `@claude @codex` 加问题或评审请求 | 只读任务并发运行;结果保持独立署名。 | | 有序交接 | `@claude Write the parser, then @codex review it` | 评审者会收到已完成的结果作为参考材料。依赖失败会阻塞该交接。 | | 明确的视角 | `@claude:critic @codex:reviewer` | 目录中的角色会塑造每个回复并限制其访问权限。 | | 结构化讨论 | Brainstorm | 选择团队与讨论策略,然后跟随其轮次与综合过程。 | | 一个协调智能体 | Mysti | 由 DeepMyst 支持的协调器可以回答、使用已配置的工具,并通过权限门控进行委派。 | | 可视化迭代 | Canvas 与 Visual Test | 处理可视化产物并检查正在运行的应用;可能需要配置浏览器。 | | 项目专属实践 | Personas、skills 与 roles | 以明确的信任与访问边界复用你团队的指导。 |  并行度取决于工作性质:独立的咨询类请求可以重叠;写入者共享同一工作区,因此串行运行。使用 **then** 表达显式依赖。[指派语法与故障排查](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/MENTIONS.md)。 ## 快速上手 1. **安装 Mysti**:从 [VS Code Marketplace](https://marketplace.visualstudio.com/items?itemName=DeepMyst.mysti) 安装,或通过 **Extensions → Install from VSIX** 安装已审阅的 beta VSIX。当 2.0 BETA 发布时选择预发布通道。 2. **打开一个可信的本地工作区。** Mysti 支持 Windows、macOS 和 Linux 上的 VS Code 1.86 或更高版本。原生语音支持与 provider CLI 有各自的宿主要求。 3. **打开 Mysti 并选择你的智能体。** 设置界面会检查安装与认证状态,并提供受支持的安装、升级或登录操作。Mysti 本身使用你的 DeepMyst 账号;CLI 智能体使用各自的凭据。 4. **从一个小任务开始。** 试试 `@claude Explain the main entry point`,当两者都配置好后,再添加 `@codex` 获取第二意见。 只需配置**一个**智能体即可开始。provider 订阅、API 访问、用量费用与支持的模型各不相同。[快速上手](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/GETTING-STARTED.md) · [Provider 设置](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROVIDERS.md)。 ## 带上你偏好的智能体 | Provider 家族 | 可用适配器 | | --- | --- | | 编码智能体 | Claude Code、OpenAI Codex、Google Gemini CLI、GitHub Copilot CLI、Cline、Cursor | | 其他编码工具 | OpenClaw、OpenCode、Qwen Code、Hermes、Continue、Kimi Code | | 本地 / API 端点 | Ollama、LocalAI、OpenRouter、MiniMax | | Mysti 协调器 | DeepMyst 账号,可选配置本地执行与委派 | 界面会跟随每个适配器的能力。模型可用性、effort、Ultracode、附件、原生审批与认证均因 provider 而异。收录某个适配器并不代表每个 provider/版本/账号组合都经过实机测试。[兼容性与设置](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROVIDERS.md)。 ## 你的上下文,你的控制权 - **上下文各归其位:** 提及工作区文件、用 `@problems` 检查诊断信息,或用 `@git` 添加 Git 摘要。 - **可见的权限:** 只读与审批模式、智能体专属的工具活动,以及跨活跃指派的取消操作。 - **连续性:** 会话历史、上下文压缩,以及针对受支持的本地变更的工作区检查点。 - **显式连接:** 选择已连接功能所使用的 DeepMyst 工具与来源。云端职责在后端持续运行;本地 Git 监视需要编辑器保持运行。 provider 请求会把提示与所提供的上下文发送给已配置的 provider。DeepMyst 功能使用你已连接的账号。遥测遵循 VS Code 的遥测设置;参见[安全策略](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/.github/SECURITY.md)与各功能专属指南。本地检查点无法撤销外部操作。 ## 浏览文档 | 指南 | 涵盖内容 | | --- | --- | | [从这里开始](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/GETTING-STARTED.md) | 安装、配置、发送第一个任务、故障排查 | | [智能体指派](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/MENTIONS.md) | 标签、角色、并行评审、依赖式交接 | | [Provider](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROVIDERS.md) | 安装、认证、模型、能力差异 | | [功能指南](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/FEATURES.md) | 聊天、菜单、听写、Canvas、权限等 | | [Brainstorm](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/BRAINSTORM.md) | 团队讨论与综合 | | [Personas 与 skills](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PERSONAS-AND-SKILLS.md) | 可复用的项目实践 | | [Proactive](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/PROACTIVE.md) | 职责、证据、通知与覆盖范围限制 | | [架构](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/ARCHITECTURE.md) | 宿主、webview、路由与 provider 边界 | | [发布说明](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/docs/releases/2.0-beta/RELEASE-NOTES.md) | 新增内容、迁移、已知限制 | | [参与贡献](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/CONTRIBUTING.md) | 开发、有意义的测试、演示与发布工作 | **翻译版本:** [العربية](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ar.md) · [中文](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.zh-CN.md) · [日本語](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ja.md) · [한국어](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ko.md) · [Español](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.es.md) · [Português](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.pt-BR.md) · [Deutsch](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.de.md) · [Français](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.fr.md) · [Türkçe](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.tr.md) · [Русский](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/README.ru.md)。社区翻译可能描述的是更早的版本;本 README 是 2.0 BETA 的当前参考。 ## 与我们共同构建 Mysti 采用 Apache-2.0 许可。有价值的贡献包括可复现的路由缺陷、原生 provider 兼容性检查、无障碍 UI 改进,以及经过测试的文档更正。 ```sh npm ci npm run typecheck npm test npm run watch ``` 使用 `.nvmrc` 中的 Node 版本;按 **F5** 启动扩展开发宿主。provider 契约、浏览器测试与贡献期望参见 [CONTRIBUTING.md](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/CONTRIBUTING.md)。 [报告缺陷](https://github.com/DeepMyst/Mysti/issues/new/choose) · [讨论想法](https://github.com/DeepMyst/Mysti/issues) · [查看路线图](https://github.com/DeepMyst/Mysti/blob/0df2961a227d53308093894d586944e9f656a0a8/plans/README.md) ## 贡献者 感谢所有帮助 Mysti 变得更好的人!