rudder

# Rudder Docs

> Canonical public documentation for Rudder, generated from scripts/docs-content-map.yml.

Rudder gives humans and agents a shared place to move work from a goal to a reviewable result. Choose Chat for conversational work or an Issue when named ownership, dependencies, or a review path help.

## English

### Start here
- [Rudder](https://docs.rudderhq.dev): Build your self-improving Agent Team.

### Get started
- [Install Rudder](https://docs.rudderhq.dev/get-started/installation): Install the Rudder desktop app and open your local workspace.
- [Run Your First Task](https://docs.rudderhq.dev/get-started/first-organization): Create a Rudder organization, connect an agent runtime, and get your first useful result in Chat.

### Concepts
- [How Rudder Works](https://docs.rudderhq.dev/concepts/overview): Learn the five ideas you need to give Agents work and keep the result useful.
- [Organize Work with Goals and Projects](https://docs.rudderhq.dev/concepts/goals-projects-issues): Use goals for direction, projects for shared context, and Issues for work that needs explicit coordination.
- [Agents, Runs, and Runtimes](https://docs.rudderhq.dev/concepts/agents): Understand the worker you configure, the tool that executes, and the record created each time it works.
- [App Builder](https://docs.rudderhq.dev/concepts/app-builder): Create or improve a local website and use it as a Rudder App on your machine. App Builder is currently in beta.
- [Built-in Browser](https://docs.rudderhq.dev/concepts/built-in-browser): Open websites beside your work and, when allowed, let a local Agent use its own controlled tabs.
- [Issues](https://docs.rudderhq.dev/concepts/issues): Use an Issue when a task needs a clear owner, status, dependencies, or review.
- [Chat, Issues, and Messenger](https://docs.rudderhq.dev/concepts/chat-messenger): Choose the right place to do a task, then use Messenger to find work waiting for your attention.
- [Calendar](https://docs.rudderhq.dev/concepts/calendar): See when Agent work ran, what is scheduled next, and where a human checkpoint happened.
- [Files, Library, and Workspaces](https://docs.rudderhq.dev/concepts/workspaces): Know where to put source material, project folders, Agent changes, and finished output.
- [Skills](https://docs.rudderhq.dev/concepts/skills): Turn a proven, repeatable way of working into instructions an Agent can reuse.
- [Automations](https://docs.rudderhq.dev/concepts/automations): Run a stable Agent task on a schedule, from a webhook, or when you trigger it yourself.
- [Review, Feedback, and Improvement](https://docs.rudderhq.dev/concepts/reviews-feedback-learning): Judge a result, explain what should change, and reuse proven lessons in future Agent work.
- [Plugins](https://docs.rudderhq.dev/concepts/plugins): Install Codex-compatible packages that combine Skills, MCP servers, and Apps.

### How-to guides
- [Use an Issue from Start to Finish](https://docs.rudderhq.dev/how-to/issue-lifecycle): Create a clear task, assign one owner, review the result, and close it with the important context intact.
- [Create an Agent](https://docs.rudderhq.dev/how-to/create-agent): Add a reusable Agent with a clear job, runtime, instructions, and safe first task.
- [Create an Automation](https://docs.rudderhq.dev/how-to/create-automation): Schedule a repeatable Agent task, choose where its result goes, and test it once.
- [Review Agent Work](https://docs.rudderhq.dev/how-to/review-agent-work): Open the real result, compare it with the request, and record a decision the next person can understand.
- [Give an Agent Project Files and Context](https://docs.rudderhq.dev/how-to/manage-workspaces-and-library): Connect a project folder, choose the source material an Agent should receive, and keep the final result.
- [Move an Organization to Another Rudder Instance](https://docs.rudderhq.dev/how-to/export-import-organization): Export supported organization data, preview the import, and safely create a new copy.
- [Install and Manage a Plugin](https://docs.rudderhq.dev/how-to/manage-plugins): Inspect a Codex Plugin package, install its supported capabilities, and configure Skills or MCP access.
- [Connect an Agent to Feishu or Lark](https://docs.rudderhq.dev/how-to/configure-feishu-integration): Authorize a Feishu or Lark bot, send a test message, and find the same conversation in Rudder.
- [Configure an Agent Runtime](https://docs.rudderhq.dev/how-to/configure-agent-runtime): Connect the program that performs an Agent's work, test it, and verify one real task.

### Reference
- [Issue Statuses](https://docs.rudderhq.dev/reference/issue-statuses): Quickly choose the right Issue status and understand what happens during review.
- [Runtime Types](https://docs.rudderhq.dev/reference/runtime-types): Compare the local programs Rudder can use to execute Agent work.
- [Where Rudder Stores and Uses Files](https://docs.rudderhq.dev/reference/workspace-boundaries): Look up the right location for project files, reusable resources, Agent state, Run changes, and backups.
- [Choose Where Automation Results Go](https://docs.rudderhq.dev/reference/automation-output-routing): Compare Chat and Issue output for an Automation, including what you see when a run fails.
- [Permissions, Credentials, and Platforms](https://docs.rudderhq.dev/reference/permissions-and-platforms): Understand local trust, where credentials stay, how external connections are limited, and what differs by platform.
- [Code Signing Policy](https://docs.rudderhq.dev/reference/code-signing-policy): How Rudder builds, reviews, approves, and signs Windows Desktop release artifacts.
- [Privacy Policy](https://docs.rudderhq.dev/reference/privacy): How Rudder handles local workspace data, accounts, optional analytics, and public-site data.
- [Approvals, Budgets, Activity, and Attention](https://docs.rudderhq.dev/reference/approvals-budgets-activity): Know where to approve sensitive actions, check spending, review changes, inspect Runs, and find work waiting for you.

### Project
- [GDPval Harness Benchmark](https://docs.rudderhq.dev/benchmarks/gdpval-harness): Methodology, results, and evidence limits for Rudder's 10-task GDPval-based harness pilot.
- [Get Help and Share Feedback](https://docs.rudderhq.dev/contact): Ask a setup question, report a bug, request a feature, or find release information.
- [Changelog](https://docs.rudderhq.dev/releases): Stable Rudder release notes, upgrade notes, and GitHub Release links.

## 简体中文

### Start here
- [Rudder](https://docs.rudderhq.dev/zh): Build your self-improving Agent Team.

### Get started
- [安装 Rudder](https://docs.rudderhq.dev/zh/get-started/installation): 安装 Rudder 桌面应用，并打开你的本地工作空间。
- [完成第一个任务](https://docs.rudderhq.dev/zh/get-started/first-organization): 创建 Rudder 组织，连接 Agent 运行工具，并在 Chat 中拿到第一个有用结果。

### Concepts
- [Rudder 如何工作](https://docs.rudderhq.dev/zh/concepts/overview): 了解把任务交给 Agent 并让结果继续发挥作用所需的五个概念。
- [用目标和项目组织工作](https://docs.rudderhq.dev/zh/concepts/goals-projects-issues): 用目标说明方向，用项目汇总上下文，用 Issue 管理需要明确协作的任务。
- [Agent、运行记录和运行工具](https://docs.rudderhq.dev/zh/concepts/agents): 分清你配置的长期 Agent、真正执行工作的工具，以及每次工作留下的记录。
- [App Builder](https://docs.rudderhq.dev/zh/concepts/app-builder): 创建或完善本地网站，并把它作为 Rudder App 在当前设备上使用。App Builder 目前处于 Beta 测试阶段。
- [内置浏览器](https://docs.rudderhq.dev/zh/concepts/built-in-browser): 在工作旁边打开网页，并在允许时让本地 Agent 使用自己的受控标签页。
- [Issue（任务单）](https://docs.rudderhq.dev/zh/concepts/issues): 任务需要明确负责人、状态、依赖或评审时，使用 Issue。
- [Chat、Issue 和 Messenger](https://docs.rudderhq.dev/zh/concepts/chat-messenger): 为任务选择合适的工作方式，再用 Messenger 找到正在等你处理的事情。
- [Calendar（日历）](https://docs.rudderhq.dev/zh/concepts/calendar): 查看 Agent 何时工作、下一次任务何时开始，以及人的检查发生在什么时候。
- [文件、Library 和工作区](https://docs.rudderhq.dev/zh/concepts/workspaces): 分清参考资料、项目目录、Agent 改动和最终产物应该放在哪里。
- [技能](https://docs.rudderhq.dev/zh/concepts/skills): 把已经验证过、会反复出现的做法整理成 Agent 可以复用的指令。
- [自动化](https://docs.rudderhq.dev/zh/concepts/automations): 按日程、Webhook 或手动触发，重复运行一项已经稳定的 Agent 任务。
- [评审、反馈和改进](https://docs.rudderhq.dev/zh/concepts/reviews-feedback-learning): 判断结果，说明哪里应该改变，并把已经验证的经验用于未来的 Agent 工作。
- [插件](https://docs.rudderhq.dev/zh/concepts/plugins): 安装由技能、MCP 服务端和应用组成的 Codex 兼容能力包。

### How-to guides
- [让 Issue 从开始走到完成](https://docs.rudderhq.dev/zh/how-to/issue-lifecycle): 写清任务，指定一个负责人，评审结果，并让重要信息留在完成记录里。
- [创建 Agent](https://docs.rudderhq.dev/zh/how-to/create-agent): 添加一个职责、运行工具和指令都清楚的长期 Agent，并完成安全的第一次任务。
- [创建自动化](https://docs.rudderhq.dev/zh/how-to/create-automation): 为一项重复 Agent 任务设置日程和结果去向，并手动测试一次。
- [评审 Agent 工作](https://docs.rudderhq.dev/zh/how-to/review-agent-work): 打开真实结果，对照任务要求，并记录下一位负责人也能理解的结论。
- [给 Agent 提供项目文件和上下文](https://docs.rudderhq.dev/zh/how-to/manage-workspaces-and-library): 连接项目文件夹，选择 Agent 应该收到的来源资料，并保存最终结果。
- [把组织迁移到另一个 Rudder 实例](https://docs.rudderhq.dev/zh/how-to/export-import-organization): 导出受支持的组织数据，预览导入内容，再安全地创建一份新副本。
- [安装和管理插件](https://docs.rudderhq.dev/zh/how-to/manage-plugins): 检查 Codex 插件软件包，安装支持的能力，并配置技能或 MCP 访问。
- [把 Agent 连接到飞书或 Lark](https://docs.rudderhq.dev/zh/how-to/configure-feishu-integration): 授权飞书或 Lark 机器人，发送测试消息，再从 Rudder 找到同一段对话。
- [配置 Agent 运行工具](https://docs.rudderhq.dev/zh/how-to/configure-agent-runtime): 连接真正执行 Agent 工作的程序，完成测试，再用一件真实小任务验证。

### Reference
- [Issue 状态](https://docs.rudderhq.dev/zh/reference/issue-statuses): 快速选择正确的 Issue 状态，并理解评审时会发生什么。
- [运行工具类型](https://docs.rudderhq.dev/zh/reference/runtime-types): 比较 Rudder 可以用来执行 Agent 工作的本地程序。
- [Rudder 在哪里保存和使用文件](https://docs.rudderhq.dev/zh/reference/workspace-boundaries): 查询项目文件、可复用资料、Agent 状态、运行改动和备份应该放在哪里。
- [选择自动化结果去向](https://docs.rudderhq.dev/zh/reference/automation-output-routing): 比较自动化的 Chat 与 Issue 结果，并了解运行失败时会看到什么。
- [权限、凭据和平台](https://docs.rudderhq.dev/zh/reference/permissions-and-platforms): 了解本地信任、凭据保存位置、外部连接限制，以及不同平台的差异。
- [代码签名政策](https://docs.rudderhq.dev/zh/reference/code-signing-policy): Rudder 如何构建、审查、批准并签名 Windows Desktop 发布产物。
- [隐私政策](https://docs.rudderhq.dev/zh/reference/privacy): Rudder 如何处理本地工作空间、账号、可选分析和公共网站数据。
- [审批、预算、活动和待处理事项](https://docs.rudderhq.dev/zh/reference/approvals-budgets-activity): 找到敏感操作审批、费用、变更记录、Agent 运行和等待处理工作的正确入口。

### Project
- [GDPval 执行框架基准测试](https://docs.rudderhq.dev/zh/benchmarks/gdpval-harness): Rudder 基于 GDPval 的 10 项任务试验：方法、结果与证据边界。
- [获取帮助与提交反馈](https://docs.rudderhq.dev/zh/contact): 咨询设置、报告缺陷、提出功能建议或查找版本信息。
- [更新日志](https://docs.rudderhq.dev/zh/releases): Rudder 稳定版本的发布记录、升级说明和 GitHub Release 链接。