Token导航 LogoToken导航

Agent Skill · 研究检索

studio-agent(工作室 Agent)

studio-agent 用于查找、检索和筛选相关信息,适合在 OpenClaw 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。

studio-agent 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为需要联网;许可证为 MIT-0,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 10,915。

总安装 11,946周安装 488GitHub Stars 公开资料未说明下载 3,826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-0
版本
0.1.7

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:studio-agent(工作室 Agent)
来源仓库:https://github.com/luketalent/studio-agent
安装命令:
openclaw skills install studio-agent
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。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install studio-agent

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。

用于 ClickZetta Studio 的任务查询、项目管理和作业执行。适用于 OpenClaw 中需要列出工作区、检查项目状态或运行自动化任务时。通过 clawhub 安装,建议结合来源仓库了解 API 权限配置。使用前需确认是否具备 OAuth 授权及项目访问范围。注意评估是否会修改任务状态或触发计费相关操作。

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 OpenClaw;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 openclaw skills install studio-agent;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

    它需要需要联网,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

studio-agent 怎么安装?

在终端执行 openclaw skills install studio-agent。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

studio-agent 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。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。

studio-agent 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

studio-agent 是免费的吗?

来源标注的许可证是 MIT-0。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。

studio-agent 用的人多吗?

公开数据是累计安装 11,946 次,近一周安装 488 次,下载 3,826 次。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。

studio-agent 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

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。

先想清楚的地方

  • 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 OpenClaw,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

  • 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
studio-agent

工作室 Agent · 本页

11,946—需要联网MIT-0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

name
studio-agent
description
Use this for ClickZetta Studio requests such as querying tasks, listing workspaces, checking projects, and creating or running ClickZetta jobs from a single JDBC secret configured in OpenClaw.
metadata

ClickZetta Studio

Use this skill when the user wants OpenClaw to work with ClickZetta Studio.

Typical triggers:

  • "帮我创建个 lakehouse 任务"
  • "用 Studio Agent 帮我 ..."
  • "在 ClickZetta 里执行/创建 ..."
  • "我当前有哪些 ClickZetta 任务"
  • "帮我查 ClickZetta 任务/作业/项目/工作区"
  • "列出当前可用 workspace"
  • "切换到 workspace 101201"
  • "刷新 workspace 列表"

Default Runtime Path

If the user asks any normal ClickZetta / Studio / Lakehouse business question and the skill already has CZ_STUDIO_JDBC_URL, immediately run:

node {baseDir}/scripts/cz-agent-oneshot.mjs --input "<user_input>"

Rules:

  • Do not ask for internal connection fields such as instance id, tenant id, project id, workspace, or password first.
  • The runner resolves runtime discovery itself from the stored JDBC secret.
  • Only fall back to setup instructions when the runner returns a concrete config/runtime error.
  • For queries like "我当前有哪些clickzetta任务", "帮我查任务", "查 workspace/project", "执行 SQL", default to the one-shot runner immediately.
  • For workspace management queries like "当前 workspace 是什么", "列出 workspace", "切换到 workspace 101201", also default to the one-shot runner immediately.
  • 列出 workspace / 切换 workspace 会使用短 TTL 的 workspace 列表缓存;刷新 workspace 列表 会强制刷新。
  • For SQL execution requests, forward the user's original wording once. Do not rewrite it into follow-up variants such as "创建一个 SQL 任务", "运行刚才创建的任务", or other speculative recovery prompts.

Install then Configure (Required)

Preferred ClawHub / OpenClaw UI flow:

  • Install the skill
  • Open the Skills page
  • Paste one JDBC URL into the secret field for CZ_STUDIO_JDBC_URL

Example JDBC URL:

jdbc:clickzetta://<instance>.<api-host>/<workspace>?username=<username>&password=<password>

The skill parses this JDBC URL at runtime and extracts:

  • instanceName from the hostname prefix
  • apiGateway from the JDBC host
  • username and password from query params
  • workspace from the path

Runtime note: if this setup step is skipped, this skill should tell the user to open the Skills page and set CZ_STUDIO_JDBC_URL.

What this skill provides

  • One-step ClickZetta access from a single JDBC secret
  • Automatic runtime login and identity discovery
  • Workspace-aware task, job, project, and SQL operations
  • A cached workspace list for faster follow-up requests

Execution policy

  • Do not ask the user for CZ_* env vars up front.
  • Assume the JDBC secret is already wired via the Skills page unless runtime errors prove otherwise.
  • When runtime reports missing/invalid Studio config, tell the user to open the Skills page and set CZ_STUDIO_JDBC_URL.
  • Default path for normal user requests: run scripts/cz-agent-oneshot.mjs once and return its result.
  • When describing connection status to the user, refer to the runtime-discovered Studio connection; do not ask the user to manually provide internal CZ_AGENT_* values.
  • For requests like "帮我创建个 lakehouse 任务", immediately:

1. run one-shot runner command 2. read runner JSON output 3. if ok=true, return content; if ok=false, return concise error and stop

  • If the runner returns ok=false, do not issue additional paraphrased or workaround requests to the Studio agent in the same turn.
  • If runner returns ok=true, treat Studio as connected and continue the user workflow; do not claim Studio is down.
  • Only ask the user for config details when runtime returns a concrete missing minimal-config error.
  • Treat config-missing as a narrow case only:

- ok=false - error.code == "PROTOCOL_ERROR" - and the message explicitly says config is missing, such as missing CZ_STUDIO_JDBC_URL or other missing Studio runtime connection info

  • If error.code == "REMOTE_ERROR", do not claim JDBC is missing or suggest reconfiguration. Report the remote error directly and keep the original error message.
  • For errors like expected string or bytes-like object, got 'NoneType', explain that the request reached the remote ClickZetta agent but failed during remote processing.
  • For SQL requests that return REMOTE_ERROR, do not suggest retrying with paraphrased prompts such as 执行 SQL: ..., 创建一个 SQL 任务 ..., or 运行刚才创建的任务.
  • Do not invent alternative SQL execution workflows unless the remote agent explicitly asks for additional fields.
  • Preferred response shape for these failures:

1. state the original remote error 2. say the request reached the remote ClickZetta agent 3. stop

  • If remote asks follow-up fields (task type, SQL details, etc.), relay those follow-up questions to the user as normal.

Mandatory one-shot command for normal requests:

node {baseDir}/scripts/cz-agent-oneshot.mjs --input "<user_input>"

Runner output is one JSON object:

  • success: {"ok":true,"content":"...","conversation_id":"...","request_id":"..."}
  • failure: {"ok":false,"error":{"code":"...","message":"..."}, ...}

Runtime Discovery

Minimal required input field for ClawHub users:

  • CZ_STUDIO_JDBC_URL

Auto-discovery behavior:

  • parse jdbc:clickzetta://... from CZ_STUDIO_JDBC_URL
  • derive the Studio instance from the JDBC host
  • sign in to ClickZetta at runtime
  • resolve the current user and tenant identity
  • load the accessible workspace list and enrich project/workspace metadata
  • initialize the Studio session connection automatically
  • workspace 选择状态单独保存在本地缓存中,不污染 ~/.openclaw/clawdbot.json
  • workspace 列表和 token 分开缓存:token 跟随登录过期时间,workspace 列表默认 5 分钟 TTL
  • if workspace / workspaceId / projectId is configured, runtime uses that workspace first and ignores local switch state
  • if the configured workspace selector matches zero or multiple workspaces, validation/runtime now fails fast instead of falling back to the default workspace

Notes:

  • In ClawHub / OpenClaw UI mode, the only user-facing field is CZ_STUDIO_JDBC_URL.
  • Runtime-derived connection data are internal and must not be requested from the user during normal use.
  • The one-shot runner also supports local workspace commands:

- 当前 workspace 是什么 - 列出 workspace - 刷新 workspace 列表 - 切换到 workspace <name|workspaceId|projectId> - 恢复默认 workspace

  • If a workspace is explicitly configured, that configured workspace takes precedence over runtime switching.
  • --replace keeps config minimal and removes stale extra keys from previous setups.
  • CZ_PROJECT_ID / CZ_WORKSPACE are now best-effort enrichments, not hard requirements.

平台分布

OpenClaw

95.21%
按下载量换算3,643

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills