Token导航 LogoToken导航

Agent Skill · 研究检索

vm0(虚拟机 0)

vm0 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

vm0 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证为 MIT,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 41,004。

总安装 1,224周安装 50GitHub Stars 56下载 392核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:vm0(虚拟机 0)
来源仓库:https://github.com/vm0-ai/vm0-skills
仓库路径:skills/vm0
安装命令:
npx skills add https://github.com/vm0-ai/vm0-skills --skill vm0
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/vm0-ai/vm0-skills --skill vm0

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

vm0 用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词、任务场景快速定位候选结果。通过 npx skills add 命令从指定 GitHub 仓库安装,需结合原始 README 核验具体用法。

安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写操作。当前无更多功能说明,建议查阅来源仓库获取详细使用指南。

来源仓库内路径:skills/vm0

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/vm0-ai/vm0-skills --skill vm0;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

vm0 怎么安装?

在终端执行 npx skills add https://github.com/vm0-ai/vm0-skills --skill vm0。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

vm0 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

vm0 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

vm0 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

vm0 用的人多吗?

公开数据是累计安装 1,224 次,近一周安装 50 次,下载 392 次,GitHub 56 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

vm0 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
vm0

虚拟机 0 · 本页

1,22456接触敏感数据MIT有提醒

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

VM0 Agent Self-Management

Use this skill to inspect and update the current agent's own configuration — its connectors (skills) and instructions — via the VM0 Zero Agents API.

Official docs: https://docs.vm0.ai

When to Use

Use this skill when you need to:

  • Read the current agent's configuration (connectors, metadata)
  • Add or remove connectors (skills) from the current agent
  • Read or update the agent's instructions
  • Redeploy the agent after configuration changes

Prerequisites

The VM0_API_TOKEN and VM0_AGENT_ORG_SLUG environment variables must be set. They are automatically injected when this skill is loaded.

The base URL defaults to https://www.vm0.ai. Override with VM0_API_URL if needed.


How to Use

1. Get Agent Configuration

Retrieve the current agent's connectors and metadata.

bash -c 'curl -s -X GET "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME?org=$VM0_AGENT_ORG_SLUG" --header "Authorization: Bearer $VM0_API_TOKEN"' | jq .

Response:

{
  "name": "agent-uuid",
  "agentComposeId": "compose-id",
  "description": "Agent description",
  "displayName": "My Agent",
  "sound": null,
  "connectors": ["github", "slack"]
}

2. Update Agent Configuration

Update the agent's connectors (skills) and metadata. The connectors field is required — it replaces the full list. The displayName, description, and sound fields are optional — only provided fields are updated.

bash -c 'curl -s -X PUT "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME?org=$VM0_AGENT_ORG_SLUG" --header "Content-Type: application/json" --header "Authorization: Bearer $VM0_API_TOKEN" -d '"'"'{"connectors": ["github", "slack", "gmail"], "displayName": "My Agent"}'"'"'' | jq .

Response: Same shape as GET.

3. Get Agent Instructions

Retrieve the current agent's instructions content and filename.

bash -c 'curl -s -X GET "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME/instructions?org=$VM0_AGENT_ORG_SLUG" --header "Authorization: Bearer $VM0_API_TOKEN"' | jq .

Response:

{
  "content": "# My Agent Instructions\n\nYou are a helpful agent...",
  "filename": "agent.md"
}

4. Update Agent Instructions

Replace the agent's instructions content.

bash -c 'curl -s -X PUT "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME/instructions?org=$VM0_AGENT_ORG_SLUG" --header "Content-Type: application/json" --header "Authorization: Bearer $VM0_API_TOKEN" -d '"'"'{"content": "# Updated Instructions\n\nNew instructions here..."}'"'"'' | jq .

Response: Same shape as GET agent configuration (name, agentComposeId, connectors, etc.).


Finding Skills

Skills come from two marketplaces. Search both when looking for new capabilities.

1. skills.sh (33,700+ community skills)

curl -s "https://skills.sh/api/search?q={keyword}"

2. vm0-ai/vm0-skills (70+ curated integrations)

Browse at: https://github.com/vm0-ai/vm0-skills

Priority rule: If a skill exists in both marketplaces, prefer the vm0-ai/vm0-skills version — it is optimized for VM0 agent workflows and has consistent quality.

Connector short names

Connectors in the API use short names that correspond to skill directory names in the vm0-ai/vm0-skills repository (e.g. github, slack, gmail, notion).

Checking required credentials

After picking a skill, read its SKILL.md frontmatter to see what secrets or vars it needs:

bash -c 'curl -s "https://raw.githubusercontent.com/vm0-ai/vm0-skills/main/{skill-name}/SKILL.md"' | head -20

Look for vm0_secrets and vm0_vars in the frontmatter. These are automatically injected when the skill is loaded — no manual setup needed. If a required secret is missing, guide the user to the VM0 console to configure it.


Operation: Self Update

Usage: self update or update myself or similar

This operation enables the agent to update itself based on user requirements.

Step 1: Understand the current agent

Read the current configuration and instructions via the API:

bash -c 'curl -s -X GET "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME?org=$VM0_AGENT_ORG_SLUG" --header "Authorization: Bearer $VM0_API_TOKEN"' | jq .
bash -c 'curl -s -X GET "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME/instructions?org=$VM0_AGENT_ORG_SLUG" --header "Authorization: Bearer $VM0_API_TOKEN"' | jq .

Step 2: Understand User Intent

Ask clarifying questions to understand what the user wants to update:

  • What new operation should be added?
  • What existing operation needs modification?
  • What skills need to be added/removed?
  • What configuration changes are needed?

Step 3: Apply Changes

To update connectors (add/remove skills):

Get the current connectors list from Step 1, modify it, then PUT:

bash -c 'curl -s -X PUT "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME?org=$VM0_AGENT_ORG_SLUG" --header "Content-Type: application/json" --header "Authorization: Bearer $VM0_API_TOKEN" -d '"'"'{"connectors": ["github", "slack", "new-skill"]}'"'"'' | jq .

To update instructions:

bash -c 'curl -s -X PUT "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME/instructions?org=$VM0_AGENT_ORG_SLUG" --header "Content-Type: application/json" --header "Authorization: Bearer $VM0_API_TOKEN" -d '"'"'{"content": "# Updated Instructions\n\nNew content here..."}'"'"'' | jq .

Operation: SaaS/Service Connection

Usage: connect to Gmail, access my Dropbox, integrate with Slack, or similar

This operation helps the user connect external SaaS services to their agent.

Step 1: Check existing skills

Read the current agent configuration and check if the requested service connector is already in the connectors list:

bash -c 'curl -s -X GET "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME?org=$VM0_AGENT_ORG_SLUG" --header "Authorization: Bearer $VM0_API_TOKEN"' | jq .connectors

Step 2: If skill exists

Guide the user to complete OAuth/connector setup on the VM0 console, then stop. Use this exact closing pattern:

Visit https://app.vm0.ai/team/<agent-name> to set up the {service} connector. Click Connect to complete OAuth authorization. The required secrets will be stored automatically — no manual configuration needed. Once connected, I'll be able to access {service} in the next conversation.

Step 3: If skill not found

Search for it in the vm0-ai/vm0-skills repository:

bash -c 'curl -s "https://api.github.com/repos/vm0-ai/vm0-skills/contents"' | jq -r '.[].name' | grep -i {service-name}

Step 4: If found in vm0-skills

Add the connector to the agent via the update API. Get current connectors, append the new one, then PUT:

bash -c 'curl -s -X PUT "https://www.vm0.ai/api/zero/agents/$VM0_AGENT_NAME?org=$VM0_AGENT_ORG_SLUG" --header "Content-Type: application/json" --header "Authorization: Bearer $VM0_API_TOKEN" -d '"'"'{"connectors": ["existing-1", "existing-2", "new-service"]}'"'"'' | jq .

Then guide the user to the VM0 console using the same closing pattern from Step 2.

Step 5: If not found anywhere

Inform the user: "This service is not supported yet. Please check the vm0-ai/vm0-skills repository for available integrations or request a new skill."

Note: Available skills can be browsed at https://github.com/vm0-ai/vm0-skills


Guidelines

  1. Agent name: The current agent name is available in the VM0_AGENT_NAME environment variable.
  2. Connectors replace the full list: The PUT endpoint replaces the entire connectors list. Always read the current list first, modify it, then PUT the complete list back.
  3. Selective metadata updates: Only displayName, description, and sound fields that are explicitly included in the PUT body are updated. Omitted fields keep their current values.
  4. 422 errors: If the API returns 422, one or more connectors reference skills that are not cached on the platform. Verify the connector names are correct.
  5. Secrets are never exposed: Secret values are never returned by the API — only ${{secrets.NAME}} references are preserved.

平台分布

Claude Code

28.25%
按下载量换算111

OpenCode

24%
按下载量换算94

Antigravity

16.18%
按下载量换算63

Gemini CLI

10.5%
按下载量换算41

windsurf

6.87%
按下载量换算27

Codex

3%
按下载量换算12

安全审计

Gen Agent Trust Hub

通过

Socket

可疑

Snyk

可疑

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在提醒或未通过项,安装前请自行复核。

继续浏览同类 Skills