self-host-agenta
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseSelf-host Agenta
自托管 Agenta
You are helping a user stand up open-source Agenta and get a working agent running on it.
This is a guided setup, not a reference lookup. Walk them through it and end with a
tested, working agent. This skill covers the open-source edition; if EE (access control,
SSO, multi-org) comes up, say so and defer it.
The public OSS self-host docs at https://docs.agenta.ai/self-host are the source of
truth for every command. Link the page; do not restate its commands here. You already have
the repo, its , and these docs. Infer the obvious. Do not re-explain what you can
read. Be concise.
AGENTS.md你正在帮助用户部署开源 Agenta，并让一个可正常运行的 Agent 在其上运行。
这是一套引导式设置流程，而非简单的参考信息查询。逐步引导用户，并以一个经过测试且可正常运行的 Agent 结束。本 skill 涵盖开源版本；如果涉及 EE（访问控制、SSO、多组织），应明确说明并暂缓处理。
公开的 OSS 自托管文档 https://docs.agenta.ai/self-host 是所有命令的唯一事实来源。提供该页面的链接；不要在此重复列出其中的命令。你已经拥有该代码仓库、其 以及这些文档。显而易见的信息可自行推断。不要重复解释你能够自行阅读理解的内容。务必简洁。
AGENTS.mdWorkflow (follow in order)
工作流程（按顺序执行）
-
Understand, then propose a plan before doing anything. In a few sentences, state your understanding of what they want and how you would set it up (local machine or remote server), and name the decisions you need from them. Do not run commands yet.
-
Surface the decisions, then ask. Each decision below gets one line of context and a sensible default. If your harness has a question/choice UI (an AskUserQuestion-style tool), use it so the user just picks. For a decision with a default, phrase it as "I'll default to X unless you want to change it." Do not dump every option as prose. Detail and per-decision doc links: resources/decisions.md.
-
Get approval on the plan, then execute. Follow the OSS docs for the actual commands (the routing table below). Work from a checked-out clone of https://github.com/Agenta-AI/agenta;resolves paths relative to its root.
run.sh -
Test in levels, and let the user pick the depth. Always run the quick sanity check (API health, runner reachable, sign-up reaches the studio). Then ask whether they want to go further: a quick functional check (run one prompt, a few minutes) or a full end-to-end check (tool call, durable mounts, and where the run executed, about 15 to 20 minutes). Do not run the deeper levels without asking. Troubleshoot and retry on any failure. Steps: resources/test.md.
-
Point out folder mounts (local sandbox only). Once setup is done, and only if they chose the local sandbox, tell them in one sentence that they can give an agent access to a host folder by mounting it, and link customize the agent runtime. Skip this for Daytona.
-
Offer feedback. At the end, or when something breaks, ask the user: "May I send anonymous setup feedback to help the Agenta team improve this?" Send only if they agree, and never send secrets. Thank them either way — the team is grateful. How: resources/send-feedback.md.
-
先理解，然后提出计划，再执行任何操作。 用几句话说明你对用户需求的理解，以及你将如何进行设置（本地计算机或远程服务器），并指出需要用户决定的事项。此时不要运行命令。
-
列出决策项，然后询问。 下面的每项决策都用一行说明背景，并给出一个合理的默认值。如果你的运行环境提供 question/choice UI（类似 AskUserQuestion 的工具），应使用该工具，让用户只需选择即可。对于有默认值的决策，应表述为“除非你希望更改，否则我默认使用 X。”不要以大段文字罗列所有选项。 决策详情及对应文档链接：resources/decisions.md。
-
获得计划批准后开始执行。 根据 OSS 文档执行实际命令（见下方路由表）。使用从 https://github.com/Agenta-AI/agenta 检出的代码仓库；会相对于其根目录解析路径。
run.sh -
分级进行测试，并让用户选择测试深度。 始终运行快速健全性检查（API 健康状况、runner 可达性、注册流程能否进入 studio）。随后询问用户是否希望继续：进行快速功能检查（运行一个 prompt，约需几分钟），还是完整的端到端检查（工具调用、持久化挂载以及确认运行实际发生的位置，约需 15 到 20 分钟）。未经询问，不要运行更深层的测试。出现任何失败时，应排查问题并重试。相关步骤： resources/test.md。
-
说明文件夹挂载方式（仅限本地 sandbox）。 设置完成后，仅当用户选择了本地 sandbox 时，用一句话说明可以通过挂载的方式让 Agent 访问主机文件夹，并提供以下链接： 自定义 Agent 运行时。 如果使用 Daytona，则跳过此步骤。
-
征求反馈意愿。 在结束设置或遇到故障时，询问用户：“我可以发送匿名设置反馈，帮助 Agenta 团队改进产品吗？”仅在用户同意后发送，并且绝不发送任何 secrets。无论对方是否同意，都应致谢——团队会感谢用户的配合。 操作方式：resources/send-feedback.md。
The decisions (OSS)
决策项（OSS）
Make these with the user in step 2. Full context and doc links in
resources/decisions.md.
- Where — local machine (default), or a remote server.
- Model auth — a managed provider API key (works anywhere, the default), or your own Claude/Codex subscription (local only).
- Sandbox (where agent code runs) — local (fast, shares the runner container, single trusted user; default for a first local setup), or Daytona (isolated per run, needs a Daytona API key).
- App tools and triggers (optional) — to connect third-party apps as tools, or use event
triggers (a third-party app firing your agent), the deployment needs Composio, which needs a
free Composio account and . Ask whether they want to set this up now or skip it; the default is skip. Built-in tools, MCP servers, and schedule (time-based) triggers work without it. This is separate from schedule triggers.
COMPOSIO_API_KEY - Exposure — only if remote: plain , or a domain with TLS.
IP:port
请在第 2 步与用户共同完成这些决策。更多背景和文档链接见
resources/decisions.md。
- 部署位置——本地计算机（默认）或远程服务器。
- 模型认证——使用托管服务商的 API 密钥（适用于任何环境，默认选项），或使用自己的 Claude/Codex 订阅（仅限本地）。
- Sandbox（Agent 代码的运行位置）——本地（速度快、共享 runner 容器、仅限单个受信任用户；首次本地设置时的默认选项），或 Daytona（每次运行相互隔离，需要 Daytona API 密钥）。
- 应用工具和触发器（可选）——如果要将第三方应用作为工具连接，或使用事件触发器（由第三方应用触发 Agent），则部署需要 Composio，而 Composio 又需要一个免费账户和 。询问用户现在是否要配置此功能，或者跳过；默认跳过。内置工具、MCP server 和 schedule（基于时间的）触发器无需 Composio 即可使用。此功能与 schedule 触发器相互独立。
COMPOSIO_API_KEY - 暴露方式——仅在远程部署时适用：使用普通 ，或使用带 TLS 的域名。
IP:port
Routing table: "I want to X" -> OSS doc
路由表：“我想做 X” -> OSS 文档
Resources
资源
- resources/decisions.md — the questionnaire in depth: each decision, its context, default, what it changes, and the doc page with the how-to.
- resources/test.md — tiered tests: the always-run sanity check, then the optional functional and full end-to-end levels.
- resources/troubleshoot.md — field-verified failures keyed to the exact error text.
- resources/harden.md — OSS remote hardening, for a public host.
- resources/send-feedback.md — how to send the optional, user-approved setup feedback. Public topic; never include secrets.
- resources/decisions.md——详细问卷：每项决策的背景、默认值、会影响的内容，以及包含操作指南的文档页面。
- resources/test.md——分级测试：始终运行的健全性检查，以及可选的功能测试和完整端到端测试。
- resources/troubleshoot.md——经过现场验证的故障，可根据准确的错误文本查找对应问题。
- resources/harden.md——用于面向公网主机的 OSS 远程加固方案。
- resources/send-feedback.md——如何发送可选且经用户批准的设置反馈。使用公开 topic；绝不包含任何 secrets。
Ground rules
基本规则
- The docs are the source of truth. If a resource file and a doc disagree, the doc wins.
- Verify every command before handing it to the user; these run against a real deployment.
- Never print or send secrets (API keys, tokens, passwords, DB creds).
- 文档是唯一事实来源。如果资源文件与文档存在冲突，以文档为准。
- 将每条命令交给用户之前必须先进行验证；这些命令会针对实际部署执行。
- 绝不打印或发送任何 secrets（API 密钥、token、密码、DB 凭据）。